I recently saw a Veritasium video on how we aren't able to measure the speed of light in one direction - only in two. He claims that there's no way to know whether light travels at what we think of as "c" in both directions or at c/2 in one direction and instantaneously in the other direction. I'm not understanding why this is impossible. I can think of a few ways that it seems possible, so I would be grateful to anyone who explains why they couldn't work.

Idea 1:

Have light travel through different media in opposite directions. For example, set up a laser to emit light through a fiber optic cable in one direction then have it hit a mirror at 45 degrees that sends it up then another at 45 degrees that sends it back, although this time through air. We'll say we have a distance of 6E8 meters to make the numbers easy. If light travels at the "standard" conception of c, that would be 2E8 going away from us through the glass and 3E8 coming back through the vaccuum. So that's 3 seconds going and 2 seconds returning, for a total of 5 seconds. But if light travels at c/2 going away and instantaneously coming back, that would be a round trip time of 6 seconds. If it were instantaneous when going away and c/2 coming back that would be a round trip of 4 seconds. I'm using the extreme difference to make the numbers easy, but even if the difference in the directions were slight this should show the difference.

Idea 2:

Use light in one direction but some other phenomenon in the other direction. For example, have a sensor that when it gets hit by light emits an audible alarm. So you have light traveling one way and sound traveling the other. In practice, this would be tricky because the speed of sound varies with atmospheric conditions and is so much slower, that the overall time would be dominated by the time it takes the sound to travel, but I would think in theory this should work. If you didn't want to use sound, you could launch a rocket or projectile back, start a conveyor belt, etc. I think these are less practical but, in theory, I don't see why they are impossible.

Idea 3:

I know the least about this one but I would think that if the speed of light were (again, to use the extreme) c/2 in one direction and instantaneous in the other, wouldn't astronomers have noticed? Wouldn't the universe look different to us? We would see the ancient universe in one direction and the "present-day" universe in the other direction. This seems very different than uniform CMB radiation everywhere. Wouldn't black holes be different if light were instantaneous in one direction? I could go on and on here because it seems like there would be so many observable differences.

  • $\begingroup$ Your first two are round trip measurements, not one way measurements. The added complications you add in both cases don’t change that. $\endgroup$
    – Dale
    Nov 7, 2020 at 22:15

1 Answer 1


This isn't a reasonable request to submit.

The point that Veritasium is raising used to be a valid point, around the year 1900. In effect it ceased to be a valid point when relativistic physics was acknowledged as the way forward.

Let me make a comparison:
There was a time that there was a theory of heat, called the theory of 'Caloric'.

Caloric was thought of as a substance that diffuses from material to material, conferring the property of heat. Caloric theory looked very strong. The higher the temperaturature difference the faster the rate of diffusion of Caloric. When Caloric enters a material it spreads everywhere, and eventually material reaches an even temperature thoughout. the fit between theory and experiment was good. In order for the equations to be that good of a fit Caloric had to be a conserved quantity. That is, Caloric theory implies there must be a law of conservation of Caloric.

Over time Caloric theory became untenable. And eventually Caloric theory was obsoleted entirely by the kinetic theory of heat.

The kinetic theory of heat comes with the message: "Look: there was never any Caloric in the first place, it only looked as if Caloric existed."

Once it has become clear that the kinetic theory of heat is the way forward it would be absurd if someone would raise the issue: "Here is an amazing thing: no chemist has ever succeeded in isolating pure Caloric. Chemists have isolated all elements of the periodic table, yet: still no pure Caloric!"

And the response to that is: "What are you talking about? There is no need to assume the existence of Caloric in the first place!"

Here is how that is an analogy for the concept of the speed of light.

The shift from newtonian mechanics to relativistic mechanics was that in terms of relativistic mechanics the metric of spacetime is the Minkowski metric.

$$ (ds)^2 = (dt)^2 - ((dx)^2 + (dy)^2 + (dz)^2) $$

This quadratic expression has strong parallels with pythagoras's theorem

If we simplify down to one spatial dimension the expression is:

$$ (ds)^2 = (dt)^2 - (dx)^2 $$

Professor Andrew Hamilton has a website where the concepts are illustrated with animations. I recommend you step in where the concept of the centre of the lightcone is introduced.

The animation depicts the shifts of plane of simultaneity as the perspective moves from one viewpoint to another, and back again.

The lightcone animations (and subsequent animations) illustrate that the lightcone is an invariant of Minkowski spacetime.

What that means is the following: before the introduction of the concept of Minkowski spacetime it was still possible to reasonably expect that there would be an underlying structure, bafflingly inaccessible to observation, and that light propagates through that underlying structure with a particular velocity. If an apparatus designed to perform a one-way measurement of the speed of light has a velocity with respect to that underlying structure then multiple processes must be conspiring to hide that velocity-with-respect-to-the-underlying-structure.

If you move to the concept of Minkowski spacetime then all of those conspiring-to-hide suppositions are no longer needed. If spacetime itself is Minkowski spacetime then all processes occurring in that arena proceed according to the geometry of Minkowski spacetime.

With the above in place I can now turn to the issue that Veritasium is raising. Veritasium:: "Here is an amazing thing: no experiment has been achieved that measures the propagation of light through the underlying structure!"

(The point is, if there is an underlying structure, and your one-way-lightspeed measuring device has a velocity with respect to the underlying structure, you will find that the speed of light is not the same in all directions.)

And ever since the adoption of relativistic physices, which was around 1910, the reply to that has been: "What are you talking about? There is no need to assume the existence of an underlying structure in the first place!"

The point that Veritasium is raising in that video is outdated - by a 100 years.

(This doesn't necessarily mean that it is inherently impossible to perform a one-way measurement of the speed of light. The point is, even if you do perform such a measurement, the outcome won't be a surprise; it will tell you something that you already know: the speed of light.)

Later edit in response to comment by NDewolf:

There is simplification in that Veritasium video, but I wouldn't rate it as oversimplification. I get the impression from the video that Derek believes he has made a significant discovery. But the point he is raising has been raised before, and the way Derek treats it is clumsy. (Obvously Derek believes his treatment is sharp.)

The way this issue should be approached is to examine the question that the physics community was faced with after the introduction of the concept of Minkowski spacetime: "Should we continue to think in terms of he way that Lorentz accounted for the null results of the Michelson-Morley experiment and the Trouton-Noble experiment, or should we move to the concept of Minkowski spacetime?"

If you move to the concept of Minkowski spacetime the issue of one way speed of light becomes moot, the speed of light is an invariant.

I concur that the issue of Lorentz/Poincaré approach versus Minkowski spacetime is still an issue worth examining, if only to know exactly what it is you are choosing for.

About the video:
In terms of level of being wrong in an educational video: it's one thing to give the wrong answer. Worse than that is to ask the wrong question.

About the task of keeping clocks in sync:
Around the world there are multiple centers for time keeping, each operating multiple atomic clocks. These centers maintain a single worldwide coordinated time: UTC The procedures to maintain that synchronized time do not involve Einstein synchronization procedure, that wouldn't work. The engineers in charge of that work know what they are doing; they make sure that a single coordinated time is available, worldwide, to nanosecond precision. Example: astronomical observatories rely on that timekeeping for worldwide coordinatation of observations.

  • $\begingroup$ I suppose you could say that Caloric theory was "obsoleted" in the sense that it stopped being mainstream. However, Claudius showed that with the principle of conservation of energy Caloric theory is equivalent to the kinetic theory of gas. So it is subsumed by the kinetic theory, atomic theory, etc... $\endgroup$ Nov 7, 2020 at 22:10
  • $\begingroup$ That said, I don't disagree with your main point. And I think caloric could be considered an "effective theory" en.wikipedia.org/wiki/Effective_theory $\endgroup$ Nov 7, 2020 at 22:15
  • $\begingroup$ @N.Steinle As widely pointed out: Caloric theory cannot accommodate that with friction any amount of heat can be produced, you keep supplying mechanical energy. But yeah, the mathematics of conduction of heat held up. Presumably that mathematics transferred readily to the kinetic theory of heat $\endgroup$
    – Cleonis
    Nov 7, 2020 at 22:35
  • 1
    $\begingroup$ I dont fully agree with the answer. The fact that some rather important works on the matter (be it physics or philosophy) were made in the last 50 years, together with the fact that there has been no definite result if there are possible physical implications, does mean that it makes sense to ask questions about it. (Notwithstanding the fact that the veritasium video might have oversimplified the discussion.) $\endgroup$
    – NDewolf
    Nov 8, 2020 at 9:57
  • 1
    $\begingroup$ Even if the the equivalence would have been rigorously proved, i think the OPs question is not that unreasonable. He isn't talking about some crackpot ideas, he is just wondering where he might have gone wrong. $\endgroup$
    – NDewolf
    Nov 8, 2020 at 10:08

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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