When Maxwell's equations are solved, one of the solutions is electromagnetic waves that should move at a certain speed ($c = \frac{1}{\sqrt{\mu_0 \epsilon_0}}$). Now, one could argue that since Maxwell's equations hold for all observers regardless of their reference frame, they should all see these waves with speed $c$. So, the speed of these waves must be independent of your reference frame.

Moreover, let's say you could observe someone travelling at more than this speed of electromagnetic waves. Now, someone in this reference frame produced these waves. Also, there is a wall in front of them that is destroyed if the electromagnetic wave touches it (it is a powerful laser). And if the person in this reference frame hits the wall, he dies. Now from your perspective, he will die since he is travelling faster than the wave and will reach it first. However, the person himself would be sure he could survive if he fired the electromagnetic wave at the wall before he got to it. So, he would be sure he would not die.

This leads to a contradiction and so it must be impossible to observe a reference frame that travels faster than this speed of the electromagnetic waves.

I suspect there is probably a hole in this line of argument but I can't imagine what it is.

  • 4
    $\begingroup$ That someone else was Einstein of course. The fact it took 50 years was because these things are far less obvious than hindsight suggests. $\endgroup$ Oct 8, 2017 at 4:25
  • 1
    $\begingroup$ So is the argument in the question sound? There are no holes in it? $\endgroup$ Oct 8, 2017 at 4:26
  • 1
    $\begingroup$ You appear to be asking why someone didn't invent special relativity in 1862. If so, that's a question about history not physics and you should post it on the HSM Stack Exchange. $\endgroup$ Oct 8, 2017 at 4:26
  • 8
    $\begingroup$ You can derive equations that give you the speed of sound in fluids and solids. These equation also hold for all observers in all reference frames (in the context of Galilean relativity). But such waves don't give rise to the surprises in SR. The 19th century physicists assumed they would deal with light in the same way that they had dealt with all the other wave phenomena they had met up to that time. $\endgroup$ Oct 8, 2017 at 4:31
  • 3
    $\begingroup$ @RohitPandey The point is that you can't know that the speed of sound waves is with respect to a medium and the speed of light isn't without experimental input. By assuming the latter, you've basically assumed that special relativity is true, so you shouldn't be surprised to see that your conclusions match special relativity. $\endgroup$
    – knzhou
    Oct 8, 2017 at 11:05

4 Answers 4


To answer your title question, definitely not. (Note: The title question has since been updated. The original was - "is it possible to deduce the constancy of the speed of light from Maxwell's equations?")

If you assume Maxwell's equations hold for all inertial frames as you are doing in your argument, then you are begging the question. You are making an assumption further to Maxwell's equations. That assumption cannot be deduced from Maxwell's equations themselves - you then need to ask how Maxwell's equations transform between different reference frames and that answer is outside the scope of Maxwell's theory.

In the 19th century, as hinted by dmckee's comment:

You can derive equations that give you the speed of sound in fluids and solids. These equation also hold for all observers in all reference frames (in the context of Galilean relativity). But such waves don't give rise to the surprises in SR. The 19th century physicists assumed they would deal with light in the same way that they had dealt with all the other wave phenomena they had met up to that time.

people simply assumed that Maxwell's equations held for an observer who was stationary with respect to the luminiferous aether, and that this aether would behave just as gas or the wood of a violin's sounding board for sound. Maxwell's equations would then change their form under the Galilean transformation. That was just "what waves did" to the mind of a physicist in 1862.

Physicists also assumed that Galileo's postulate, that only relative motion between observers were experimentally detectable, didn't hold for electromagnetism. Galileo, after all, knew nothing of electromagnetism.

And that is a perfectly plausible theory to postulate: from a logic standpoint, it is just as valid as Einstein's second postulate and the restoration of Galileo's relativity principle. These things cannot be settled by logic, but only by asking Nature for her take on these things. That is, they can only be settled by experiment - and we find that Maxwell's equations do indeed transform covariantly and that the transformation between relatively moving inertial frames is the Lorentz, not Galilee, transformation.

  • $\begingroup$ It is clear that observers in two different reference frames will correctly come to the conclusion that Maxwell's equations hold for them (weather or not they are aware of each other) and hence that the speed of electromagnetic waves for them is $c$. When you say, "you then need to ask how Maxwell's equations transform", do you mean "sure, Maxwell's equations hold in my reference frame, but will they hold in that other persons reference frame as seen by me"? $\endgroup$ Oct 8, 2017 at 6:19
  • $\begingroup$ If so, I think my argument doesn't rely on this. As long as the two people see the same speed of light in their own reference frames, the contradiction in my argument seems to stand. $\endgroup$ Oct 8, 2017 at 6:25
  • $\begingroup$ Also, to your point of not relying on logic to settle these things, Einstein made many arguments and predictions that were essentially thought experiments. These were verified by real world experiments only decades later. Not saying of course that my argument is similar to the ones he made (it might have a flaw, I just haven't seen it yet). $\endgroup$ Oct 8, 2017 at 6:37
  • 2
    $\begingroup$ @RohitPandey Your assumption of the constancy of the speed of light is a further assumption about how Maxwell's equations transform. Also, of course Einstein made arguments, beginning from postulates, i.e. that Galileo's relativity principle held good and that $c$ is the same in all inertial frames. The luminiferous aether theory began from different postulates - that light had a medium and that the transformation undergone by Maxwell's equations was described by a Galilean boost. There is no way to tell from logic alone which of those theories is correct: only experiment can settle that ... $\endgroup$ Oct 8, 2017 at 7:05
  • 2
    $\begingroup$ @RohitPandey ...question. In that sense, Einstein was just plain lucky (not to detract from him, but there is an element of luck involved). Sure (and I think of myself as a theoretician), we can and do seek insight from similar behaviors, symmetries, analogies. These help us economize on our experimental budgets. But that is all that theory does for us: questions such as you ask cannot be inferred by logic alone. That situation is what sunders mathematics and logic from physics. $\endgroup$ Oct 8, 2017 at 7:07

Do Maxwells equations imply the constancy of the speed of light?

Maxwell's equations are a rigorous mathematical system and yes, it can be shown that they can have a Lorentz covariant form, after some mathematical transformations.

Then there are the Lorentz transformations themselves, that were necessary to explain the results of the Michelson Morley experiment, no luminiferous aether, in 1881.

He or someone else should have reached the conclusion that nothing can travel faster than light shortly after.

It takes some time for physics to shift a paradigm,like the existence of luminiferous aether, which was a basic belief since ancient times, and it was not before special relativity that things fell into place.

Progress in physics is incremental and dependent on mathematics.

I do think your gedanken experiment is fine. At the period of the discovery of Maxwell's equation, there was no fashion of using gedanken experiments. The Reductio ad absurdum was not used in physics, which had recently ( 100 years?)discovered the power of calculus and differential equations. Until the Michelson Morley experiment the ether was a very real basis for the physics at that time, it was a medium through which everything traveled.

We learned to play with thought experiments due to special relativity and quantum mechanics. At that time there was not even a dream that light could be destructive, for example, lasers were not in that time frame of reference. That is what I mean by paradigm shift.

  • $\begingroup$ yes, it is clear from the first PDF that you linked that Maxwell's equations have a Lorenz covariant form. However, it took 7-8 pages and some equations. If you trust the argument in my question, it seems to have achieved the same in one paragraph. This strongly suggests that the argument is flawed. The question is, what is the flaw. $\endgroup$ Oct 8, 2017 at 5:18
  • $\begingroup$ @RobJeffries corrected $\endgroup$
    – anna v
    Oct 8, 2017 at 11:06

The speed of light of has been reduced to zero. In your argument, an observer, with any positive speed, firing a monochromatic probe beam of unlucky frequency through this medium to detect the wall will hit the wall before the beam does. Your argument then concludes that any positive speed for the observer is impossible.

Nothing in Maxwell's equations fixes the speed of light. As soon as we observe two different speeds of light, we have two different speed limits in your argument. Which one is the real speed limit?

It is a triviality to observe matter and light traveling at various speeds, including matter apparently exceeding the speed (more precisely, phase velocity) of light in a medium (Cherenkov radiation, linked below). You write "one could argue that [... Maxwell's equations satisfy special relativity ...]". You assume Special Relativity at this point in your argument. Of course, if you assume SR, don't be surprised when you get SR. As counterpoint, what makes you so sure Maxwell's equations aren't an effective theory, only valid for low intensities, like Newton's gravitation? or large distances (more than, say, ten wavelengths)? or where the luminiferous ether is particularly rarefied?

When you write "... they should all see these waves with speed $c$." I believe you mean $c_{\omega,\theta}(O, \Lambda)$, where $\omega$ tracks the frequency dependence (see dispersion), $\theta$ is the polarization (see birefringence), and $O$ and $\Lambda$ record (at least data on the zeroeth, first, and second derivatives of) the reference frames of the observer and the light interacting with matter (see Cherenkov radiation) relative to the luminiferous ether. And there may well be additional parameters we need to track, like intensity. Since your argument becomes (at least) ambiguous as soon as one has observations of different speeds of light, all of these cause trouble for your assumption of SR.

  • 1
    $\begingroup$ The argument is fine when the speed is measured with respect to a medium. The contradiction doesn't hold anymore since the person moving towards the wall shouldn't complain if he's moving faster through the medium than the speed of light in that medium. But yeah, when we say "free space", how do we know it isn't just another medium? That is the key point. $\endgroup$ Oct 8, 2017 at 20:58
  • 1
    $\begingroup$ @RohitPandey : There's a reason I keep mentioning Luminiferous ether. Maxwell believed in it. He also believed it would produce Galilean (not Lorentzian) boosts. The expected result of Michelson-Morley was a location varying speed of light, depending on the motion of the Earth with respect to the ether. Assuming a constant speed of light for all observers is very much 20/20 SR hindsight, an anachronism to the context of Maxwell's equations. $\endgroup$ Oct 8, 2017 at 22:59

I think the hole in the argument is the affermation that Maxwell's equations hold true for all reference frames.

AFAIK they don't.

They hold true in what Einstein referred to as a local reality.

  • $\begingroup$ Imagine I on Earth, derive Maxwell's equations with all its parameters. I then phone my friend on Mars and tell him the equations and my approach. Are you saying the friend of Mars will not reach the same four equations? $\endgroup$ Oct 8, 2017 at 20:53
  • $\begingroup$ A local reality does not exclude there may be other identical local realities. Or that they may account for 99% of the instances. But if your friend was on the image horizon of a black hole would he reach the same equations? $\endgroup$ Oct 10, 2017 at 6:19

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