Is there any way (practical or theoretical) to measure the one-way speed of light?

The two methods that come to mind are:

  1. Stellar aberration, and

  2. Using adiabatic clocks: synchronize clocks, then slowly move them apart

I think 2 is not really measuring one-way speed, although I can't work out exactly why.

Is 1 measuring one-way speed? In particular, if the one-way speed was not isotropic, would there be a different amount of stellar aberration if the telescope is pointed at stars that lie in exactly opposite directions?

  • 2
    $\begingroup$ Interesting question. What one could definitely measure one way with high precision is the difference between two speeds, e.g. the speed of light in vacuum and the speed of light in an optical medium. $\endgroup$ – CuriousOne Sep 7 '15 at 4:17
  • 4
  • $\begingroup$ Looks like this might get closed as a duplicate. I hope OP will write a question specifically about option 2, as I've never heard of it before. $\endgroup$ – DanielSank Sep 10 '15 at 0:57
  • $\begingroup$ Why one would need to measure a one way speed of light? What would one expect to be different from the measurement with the return in? $\endgroup$ – anna v Apr 20 '17 at 14:06
  • $\begingroup$ @annav. If someone measures one - way velocity of light and this velocity is not c, that immediately turns SR into idle fiction even for raving fans. Good news that is hardly possible and gives perfect opportunity to speculate. To measure one way speed of light one has to synchronize two spatially separated clocks, and to synchronize these clock one has to know one way speed of light. youtube.com/watch?v=9XjS4I4oQDY $\endgroup$ – Albert Apr 20 '17 at 18:49

How about using something which has well-defined speed but not at the speed of light? As an extreme example, suppose you have a conveyor belt, marked at regular intervals, and very well-calibrated speed. Then you know how long it takes for one tick-mark to go from the starting point to where the observer is. Synchronize the output pulse of light to a tick-mark, and record the time of arrival of both the light pulse and the tickmark.

I recognize that you'll need an extremely well-calibrated belt drive (and maybe a 20km belt :-) ), but perhaps one can extend the concept to, say, speed of sound through homogeneous rock.

  • $\begingroup$ The problem is that we have to know speed of the mark on the belt. To measure this speed we need two synchronized clocks on the point of departure ant point of arrival of the mark. How to synchronize these clocks? The greatest velocity is c. We can synchronize these clocks by light, but measured velocity will be c!!! We cannot send a signal faster than c! $\endgroup$ – Albert Apr 20 '17 at 12:48
  • 1
    $\begingroup$ @Albert I don't think so. We need to know the distance between departure and arrival, but since we've placed a sequence of tickmarks on the belt, we know that in our reference frame, there is a tickmark simultaneously at departure and arrival locations. It doesn't have to be the same physical mark. But I understand your comment - my answer was not completely clear. $\endgroup$ – Carl Witthoft Apr 20 '17 at 13:16
  • 1
    $\begingroup$ I think that to say that marks are "simultaneously at departure and arrival locations" corresponding clocks in these locations have to show the same time. We run into the same problem - how to synchronize them without assumptions about velocity of light one way? Comprehensive analysis of any known way to measure one - way velocity (it seems) leads to conclusion that it is velocity back and forth. $\endgroup$ – Albert Apr 20 '17 at 13:26
  • $\begingroup$ One can use synchronized to zero at departure atomic clocks on all the ticks? $\endgroup$ – anna v Apr 22 '17 at 3:47
  • 1
    $\begingroup$ That's why it is often said - Einstein synchrony convention. Convention! Einstein ASSUMES that one way speed of light is c and synchronizes clocks in observers rest frame this way. But this is not confirmed physical measurement. Well known constant c is a speed back and forth. $\endgroup$ – Albert Apr 22 '17 at 6:49

It is hardly possible to measure one –way speed of light (to get any other value than c), because it is necessary to synchronize distant clocks first. One clock is at point of departure and another at the point of arrival. To synchronize clocks it is necessary to know that one way speed of light, since it is the “greatest available speed” and no instantaneous transfer of signal is possible.

Method Nr. 2 is known as slow clock transport and is equivalent to Einstein synchrony convention. Measured by means of this synchronization method velocity of light will be c.

Some people think that Roemer measured one - way speed of light. Actually Roemer measured speed of light back and forth, Australian physicist Karlov showed that in his analysis.

It is also not possible to synchronize clocks by means of rigid rod, since absolutely rigid bodies do not exist and signal cannot move inside the rod faster than light. As a consequence it is not possible to measure one way speed of light by synchronizing clocks by a chain (S. Marinov once proposed this synchronization in rotating mirrors experiment) which connects two clock faces. Momentum will not travel through the chain faster than light.

It is also impossible to measure one way velocity of light(other than c) by two discs that rotate with the same velocity in immediate vicinity. Then you move these discs in different directions. But, if the ether exists, due to motion through the ether these discs will slow down at different magnitude and will be finally out of sync. So, measured velocity will be c.


One of the most interesting methods was so - called Double Fizeau Toothed wheel. There are two clock faces (or toothed wheels) that are connected by rigid rood. It implies that elastic force has to keep the rod straight all the time. It seems that after initial acceleration the discs have to rotate at the same phase. An experimenter sends the beam through one wheel first and the beam goes through the next wheel then. By angular velocity (when the beam vanishes) it seems that we can calculate velocity of light one way. Some people believe that this measurement (if properly done) will allow to measure one way speed of light.

This method was employed by Stephan Marinov and M.D. Farid Ahmet.

But (this is my opinion and some other researchers) the rod will twist in laboratory (if the ether exist). That twist will compensate differences in velocities and measured velocity will be c.

Herbert Ives (who measured time dilation by means of Transverse Doppler Effect) considered this method in his 1939 article and came to the same conclusion also. Good to note that Ives never accepted SR https://www.osapublishing.org/josa/abstract.cfm?uri=josa-29-11-472

However, all these methods do not resolve the problem - whether one way velocity of light is c in all directions. Velocity of light clockwise and counterclockwise will be different on rotating ring, if measured with single clock or if the clocks are synchronized from the center. But, if we will measure velocity of light on certain segment of the rings by two Einstein – synchronized clocks (light is sent from clock to clock on the ring) measured velocity will be c.

Angle of aberration depends on velocity of light, but if we assume that one way velocity of light in ether is c, then it will be greater if you move towards source of light or lower if you recede from it. This way angle of aberration will always be the same despite of velocity of your laboratory.


Einstein synchronization is a convention that doesn't prove that velocity of light in all direction is c. Reichenbach's synchrony convention is self consistent and admits that velocity of light is different in different directions, while measured velocity back and forth is c. For example velocity of light back can be infinitely large and forth infinitely close to c/2.



Interesting book by Max Jammer:

https://www.amazon.com/Concepts-Simultaneity-Antiquity-Einstein -Beyond/dp/0801884225

In Lorentz theory velocity of light is the same in all directions only in preferred frame ( Ether). For moving observer it will be different in different directions. But, measurements of length and clock rate by moving observer and one at rest in the Ether will be the same (measuring rod contracts and clock dilates), if the moving observer cannot detect his own motion through the Ether and employs the same Einstein signaling method for clock synchronization in his frame

Chapter 3.5.5 http://www.mpiwg-berlin.mpg.de/litserv/diss/janssen_diss/Chapter3.pdf

It s possible simulate the whole Special Relativity (with all effects – Time dilation, length contraction, Barn and Ladder paradox, Twin paradox, Bell’s spaceship paradox, relativistic velocity addition, relativistic Doppler Effect e.t.c) in medium - even in a pond in your backyard.




One way speed of light measurement could be reduced to synchronizing distant clocks. Let's use Einstein's boxcar experiment with slight modification: Initially let's have a stationary boxcar and mark exactly points A, A', B, B'where A and B are on the embankment and A'and B' attached to the boxcar. at point A let's install a laser pointing directly to a mirror at A'(attached to the boxcar) which is reflecting the laser beam back to a photo sensor at A.Let's make the same arrangement at B and B'.If we move boxcar with non-relativistic speed (uniform )v from A to B (or from B to A) the mirrors at A'and B' will reflect the light at exactly the same moment, so we can synchronize the clocks at A and B. Accuracy of such synchronization would depend how precise A, A'and B and B' are marked, how big is the distance between A and B and how fast the boxcar is moving.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.