So, the idea is, let’s say it’s infinitely fast in the East->West direction, but half as fast in the West-East direction, so whether you measure it East->West and back or W-E-W, it will still measure the same.
And I guess, if you measure it N-S-N, it would be some combination of infinitely fast and half c?
Also, in the Veritasium video Crafter_Man linked to, he points out that measuring or knowing the one-way speed of light is probably fundamentally unknowable. Special relativity always factors in a round-trip measurement, for any clock-comparing equation or experiment. The very concept of a one-way speed might turn out to be meaningless. The physics always works out, no matter how you dice up the travel times.
That’s how I’m understanding it.
When I was taking Geometry in High School, our teacher happen to mention that trisecting an arbitrary angle with a protractor and a straightedge is impossible. But that did not stop me from spending countless hours trying.
All that to say, that I find puzzles like this irresistible. So here’s my ideas for measuring the one way speed of light - where did I go wrong?
Idea 1. Sync two clocks together. Move one far away, then (at the turn around spot) bring it back. There will be drift, right? Can’t you take the drift amount, divide it by 2 and assume that number (D) is the one way drift amount? So, sync two clocks again then adjust one of them by D, send that clock back to the turn around spot and they should be synced again, right? That should allow you to measure one way c.
Forget Idea 2, I just found the flaw in the logic myself.
I think the issue in your first scenario is that bringing the 2nd clock back, introduces more drift again, and just dividing the discrepancy by 2 assumes C is equal in all directions again.
Quote from the Veritasium video (it really is a good explainer):
“The reality is we’re stuck. We need synchronized clocks to measure the one-way speed of light. But we need to know the one-way speed of light in order to synchronize our clocks.”
C’mon, guys. This is so simple I don’t understand why all the physicists in the world have missed it.
Just run down the street with a stopwatch. Easy peasy.
I suspect the trouble is that in using clocks like this to try to validate isotropic light speed, you cannot avoid folding minimally SR into the experiment. Now if c is not isotropic, SR will tell us that the time dilation of the outward and inward movement of a clock must be affected by a Lorentz factor that is calculated with different values of c in each direction of travel. So the accumulated time offset is also anisotropic, and the offset exactly wipes out the measurement of the anisotropy of c.
Send one clock out, and bring it back. Yes the two clocks no long longer match. Even if you validate that the difference is entirely due to SR, and you send the other clock on the same trip, and they now match when it comes, home, you still don’t know what the offset was at the far end of the trip. Whether c is isotropic or not, you are none the wiser.
Where it gets weirder is if we are only talking about light speed being anisotropic and c remaining isotropic. That might be so in a real universe to a very very teensy tiny extent. Certainly observations of very distant events tend to rule this out. Weird properties of space have been proposed that could have caused a gap between c and light speed. Also interaction with CMB. Interaction with the CMB would just be real world correction of the properties of intergalactic space, and not actually a true anisotropic property.
Love this discussion all of you! I’m working on some ideas…
If light were not isometric. You can’t divide by 2 sorry.
But you CAN know the worst case time offset! Which ruins the precision. But I only want 1%.
I won’t discourage you from doing so, as it is a healthy and brain-stimulating exercise. But you will undoubtedly discover that there’s not a valid way of doing it, much like some people eventually realize that there’s no way to build a perpetual motion machine.
You may be right. I can actually carry out some experiments for little money.
Let us know how it goes, and report your results. (And I am not being facetious here.)
sbright33, did you watch the Veritasium video linked both here and in the other thread?
I think measuring over long distances causes a lot of problems. Get a better clock and measure over a very short distance. Just a few inches apart will minimize all those relativistic effects.
IANAP, but I don’t think the issue is excessive uncertainties at long distances. The underlying issue is cooked into the equations, and you will have this problem regardless of the distance between the sensors.
That wasn’t serious. Although I do wonder what the minimum distance for an accurate measurement would be simply to make any experiment more practical to conduct.
That seems to be the crux of the problem and what is not being understood.
Using clocks at all IS a problem. The presumption is that if you sync the clocks at your source (wherever you are), the clocks will remain ‘synced’ when you move one clock whatever distance - whether to the moon or to the next county over, accounting only for the travel distance.
But that’s what you are trying to establish in the first place. If the speed of light is not isotropic, you can’t establish they are sync’ed. You would need to bring it back to check the sync, and that makes it a two way measurement.
Having the clock report its time at one end and sending the results back by messenger pigeon (or whatever), doesn’t solve that issue. The “number of ticks” or whatever will be internally consistent with the ‘known distance’ (which itself is a fuzzy concept the way it’s being described). That is, the idea that the distance is some fixed quantity independent of the speed of light is itself a major assumption equivalent to assuming the speed of light is isotropic.
Yes I watched it many times. AntiBob we agree. We understand each other. See we are communicating successfully!
We have sent an atomic clock around the world on a commercial jet and noticed the time difference to the one at home. We can calculate this. We can send it in one direction, say East, and back. It matches the calculation. But we don’t know if this “drift” occurred going out, or coming back. If the speed of light is c/2 in one direction, the clock could either be the same at the East point, or all of the delta Time has already happened on the way out, and none on the way back. Or anywhere in between. Let’s do 2 tests assuming each of these extremes, to measure the speed of light One Way. We know dtime is not negative in either direction. After we do this test we have calculated that the speed is c/2 or really fast. But we don’t know which. This illustrates the problem?