A photon collides with an atom, the atom absorbs the energy, and then the atom emits a photon. This photon collides with *another *atom, the atom absorbs the energy, and then the atom emits a photon. And on and on.
When the photon is traveling *between *atoms, it is traveling at a velocity of c. But when you measure the average velocity of the photon over a given distance, it is slower than c, because it takes time for each atom to adsorb and emit a photon. In other words, a photon is always traveling at c regardless of the medium.
Imagine you only ever measured it in air, and only ever got the slightly slower speed, and someone asked “What’s the true speed of light, the speed between atoms?” That’s analogous to what we’re asking about. The speed of light is only ever measured in the presence of virtual particles.
This is exactly the sort of question that requires precise distinct between the physical constant c and, separately, the speed that light travels through any particular medium. The speed that light travels might change, but the speed of light (i.e., the annoyingly named constant c) needn’t.
Consider this experiment: Measure the energy released from the decay of a radioactive nucleus (calorimeter) and measure the mass difference of the nucleus before and after the decay (mass spectrometer). Divide those two numbers to obtain c[sup]2[/sup], since E=mc[sup]2[/sup], without any reference to light or the properties of the vacuum at all.
Sometimes it is possible to describe light as if it were made of photons which behave like relativistic, subatomic ‘billiard balls’, but refraction really, really doesn’t lend itself to this model of light.
So, I think we still remain unclear about the question about the speed of photons in vacuum versus c, and as ZenBeam noted, also the speed of other massless particles.
There are clearly some subtleties I had never thought about. And the wretched nomenclature does seem to get in the way of a clear discussion of the question.
So, c is 1, is what we understand as c from SR. It is invariant.
We have ls - the speed of photons in vacuum.
We might have gs - the propagation speed of gravity (perhaps the speed of a graviton).
Perhaps we can have ts - the speed of a thingyon - an as yet unknown massless particle. (Great shame it can’t be a neutrino anymore.)
Does the existence of virtual particles in the vacuum mean that ls is intrinsically ever so slightly less than c? If so, how much?
Could gs therefore be different to ls in a very subtle manner? Heck, if we ever get a theory of quantum gravitation might there be equivalent tricks in the vacuum that mean gravity propagates every so slightly less than c, or does this actually make no sense?
Thanks guys. I was aware that the question is subtle, and I agree that the tests are worthwhile; I just was surprised at how often I run across tests of this nature.
As to “c” and “C” and “speed of light”: I suppose you’re right. Disappointing, really. It would help to remove a lot of the misconceptions out there if “light” were not treated as something ultra-special.
The vacuum in SR is Minkowsky space, and in that space, light propagates at c; if you introduce the Casimir plates, you really just create a different vacuum, in which light may propagate faster. The Casimir vacuum isn’t really ‘vanilla SR’ anymore, and that you can obtain a different speed for light might not be any more shocking than the same thing happening in a non-inertial frame of reference. (But I’m not really sure about this point.)
The speed of light in (Minkowski) vacuum in QED must equal c, because you put it into the theory by construction; likewise, the speed of gravity, if gravity is described by a relativistic QFT (or some generalization true to the same principles, like string theory) must be c. That’s because all effects at spacelike separation commute, which is an assumption you include into the theory from the start.
I think that every time you use your GPS, you are implicitly measuring the speed of light. And it isn’t changing since the GPSs are still working. I once heard a talk by a physicist who said that if general relativity weren’t taken into account, the GPS would drift by close to miles per day. General relativity causes the clocks in orbit to run slightly more slowly and I imagine that is some miserable fraction of 1%. The OP claims that the speed of light is now more than 1% slower than it was in 1958. Absurd!
Incidentally, the Michaelson-Morely experiment, which I think was done about 125 years ago was carried out with an interferometer.
**UPDATE: Pasta did not actually do this deliberately – he cut and pasted from Google and did not see the offensive portion of the link.
He did nothing wrong at all. We did not understand what was going on here, now we do and we’re working on the problem.
Warning is of course rescinded.**
MODERATOR WARNING
I have edited your link to remove the offending portion.
The first rule of the Straight Dope is “don’t be a jerk.” Playing stupid games like altering quote links to display in a way that is offensive to some people is being a jerk.
Heh, I did put ether in quotes. There is vacuum energy, which has some resemblance to a medium of space, only it doesn’t really act like ether now does it?
You’re not correctly distinguishing between c and the speed of light here. The Lorentz equations use c. Light propagates at ls, to use Francis Vaughn’s notation.
You can’t just assume c = ls here, because that is what we are questioning. Perhaps the correct answer is ls = c * (1 - 1e-30), but you’re currently unable to measure that difference. Your precision isn’t good enough to distinguish between c and ls.
