There is a well known 2 mirror clock time dilation thought experiment in STR. However it is a bit unclear to me due to its unrealistic nature. By unrealistic I mean that experiment implies that stationary observer can immediately "see" the photon trace from side without photon actually hitting the detector (eye). As far as I know you cant see photon from side without it being absorbed by detector.

So I made a variation of this experiment, your answers to which, I hope would clarify some things about STR (for me at least). experiment setup

The rocket with two photon emitters S1, S2 is moving at constant speed V (close to light speed) relative to R1 photon receiver which we will consider "stationary". Outside the rocket the only source of EMR are S1, S2 photon emitters.

The questions are:

  • Will photons emitted by S1 and S2 (we consider distance between them is 0) be received at R1 simultaneously from R1's reference frame.

  • Does it matter what kind of clocks I use for S2 emitter?

  • Suppose I use same type of photon clock but horizontal (relative to movement direction), will S1 and S2 photons arrive simultaneously from R1 frame?

  • Does STR imply that all kind of clocks work the same, so if I'm viewing all possible clock types from reference frame A and they agree between themself, they all must agree when I will be viewing them from frame B, or any other frame.

  • $\begingroup$ There is not (and never has been) a requirement that anyone "see" the light pulse. The path is inferable by the fact that it leaves the emitter and arrives at the detector. $\endgroup$ Feb 1, 2018 at 1:41
  • $\begingroup$ I'm not sure I understand this, but I think you're saying that S1 and S2 are in (for all practical purposes) the same location, in which case everyone has to agree on whether their photon emissions are simultaneous. Obviously, two simultaneous emissions from the same location are going to both hit R1 (or anything else) simultaneously. Is there something else you're trying to get at? $\endgroup$
    – WillO
    Feb 1, 2018 at 6:25
  • $\begingroup$ @WillO Yes, there is something I'm trying to get at. At your answer physics.stackexchange.com/a/276603/65748 you said in bold text "even if the vertical clock is not there." which made me worried. Also 'dmckee' said "The path is inferable", which I understand as "we imagine", so do time dilation must be imaginable. So I want to hear explanation of my experiment from STR point of view. If I do analogous explanation, we "infer" that photon between mirror trace a longer path, so S1 emitter ticks slower, so S1 and S2 get out of sync and their signals ARE NOT received at R1 simultaneously $\endgroup$ Feb 1, 2018 at 9:05
  • $\begingroup$ @WillO After some discussions here on stackexchange I realized that what I wanted to know designing this experiment, is - do SRT imply that all clocks are identical. I mean all elementary particles and everything in universe "use" same universal clocks, which undergo same changes in all conditions. $\endgroup$ Feb 1, 2018 at 14:25
  • 1
    $\begingroup$ The traveler cannot infer it alone. As Will said it is "forbidden by relativity". If your frame is moving at a constant velocity and you've no outside sources to compare with then you shouldn't be able to tell you're moving at all. The sources are moving with your ship and, as such, would defy this logic. $\endgroup$ Feb 1, 2018 at 15:16

3 Answers 3


The difficult thing to overcome when first approaching special relativity is the idea of a "True" time.

The clock in S2 will be effected just as much as the results of the light bouncing between two mirrors

You won't have a difference between when the S1 and S2 emit light because, to an observer on the ship moving at a constant velocity, they are both clocks measuring the same thing. To an observer at R1 the signals from S1 and S2 will also be identical, neither will agree with a clock that was in R1's frame of reference.

  • 1
    $\begingroup$ I agree. All he's basically done is define another identical "mirror clock" for S2 (implicitly and undrawn) which is synchronised with the first mirror clock which S1 uses, so they may as well just both use the same clock. $\endgroup$
    – Steve
    Feb 1, 2018 at 16:21
  • $\begingroup$ I has overcome True time idea long ago, I disagree with idea of "True" clocks. Look at this answer (and question) physics.stackexchange.com/a/276614/65748. What if the clocks are not length based and are unaffected by length contraction. $\endgroup$ Feb 1, 2018 at 16:23
  • $\begingroup$ @Steve I didn't defined another "mirror clock" I didn't said what kind of clocks are for S2! Suppose that its clocks based on spin of electron. $\endgroup$ Feb 1, 2018 at 16:26
  • $\begingroup$ @AlexBurtsev, length contraction is not a real effect - it is more of a perceived effect when using Doppler ranging for moving objects, and it is symmetrical. Time dilation is a real effect that can be experienced asymmetrically, and it's got nothing to do with the length of the clocks, it's because the speed of light (between two relative points) changes in such a way as to introduce an isotropic delay in all directions. As for the type of clock used by S2, it doesn't seem to matter, because movement through space appears to affect all processes in a very fundamental way. $\endgroup$
    – Steve
    Feb 1, 2018 at 16:42
  • $\begingroup$ @Steve I don't want to start this cause it involves long terminology sync for what we name real. But if you think Time dilation real, you must also think that length contraction is real, en.wikipedia.org/wiki/… $\endgroup$ Feb 1, 2018 at 16:52

It gets confusing when comparing light clocks to mechanical clocks, etc. I hope this helps.

enter image description here enter image description here

As you see above, Jack is at rest, meaning he has neither accelerated no decelerated, and thus has not changed his frame of reference. Jack uses 2 synchronized clocks to measure how much slower Jill's clock is ticking. The results are also noted above.

Now imagine that Jill's rocket is very long and that there is also a clock located at the rear end of the rocket. The distance between these clocks is the same as the distance between Jack's clocks.

Jill takes off to the left, stops, and then accelerates to the same speed previously noted, and heads back to pass by Jack. From Jill's perspective, both her clocks still seem to be synchronized. However, from Jack's perspective, he sees that Jill's rear clock is ahead of the one at the front.

Now if Jill uses her two clocks to measure how fast one of Jack's clocks is ticking, the clock time offset between her two clocks will in turn have her measurements indicate that Jack's clock is ticking slower than her clocks, and do so even though Jack has not altered his frame of reference.

With this being the case, dozens of rockets that are identical to Jill's rocket, could have left jack and then returned and passed by him, but all do so at different speeds. They all, when measuring one of Jack's clocks, would in turn get different results. But there is only one Jack, not multiple Jack's.


There is a well known 2 mirror clock time dilation thought experiment in STR. However it is a bit unclear to me due to its unrealistic nature. By unrealistic I mean that experiment implies that stationary observer can immediately "see" the photon trace from side without photon actually hitting the detector (eye). As far as I know you cant see photon from side without it being absorbed by detector.

I am not sure I have understood you well, but in SR an observer does not look at the moving clock (light clock) from side. Since you call it "unrealistic" I'd try to shed some light:

In special relativity an observer is not a single person or human being, who sits on a bench and observes moving clocks. Stationary observer is the whole reference frame, or the team of observers. These observers stay apart from each other at certain distance and each has a clock. These clocks have been previously synchronized by Einstein signalling method (there is also Reichenbah's, which is self - consistent). They record space and time coordinates of events in immediate vicinity, or straight in front of each of them.


A stationary observer of reference frame $S$ conducts measurements this way: He places light clock $C_1$ at coordinate $x_1$ of his frame and light clock $C_2$ at coordinate $x_2$ of his frame.

Let these light clocks have digital oscillation counters.

Then this observer sends a beam of light from clock $C_1$ towards clock $C_2$ when clock $C_1$ shows 0. He assumes, that one - way speed of light is c (Einstein synchrony convention). Since he knows distance and speed of light, he synchronizes these clocks, so as they show the same time in reference frame $S$.


Imagine light clock $C'$ passes by clock $C_1$ first at moment of time $t_1$ and clock $C_2$ at moment of time $t_2$ then. At these moments, readings of the moving clock and the corresponding fixed clock of reference frame $S$ next to it are compared.

Fig.1 Fig.2

Let the counters of moving clock measure the time interval $\tau _ {0}$ during the movement from the point $x_ {1}$ to the point $x_ {2}$ and the counters of clocks $C_1$ and $C_2$ of the fixed or “rest” frame $S$, will measure the time interval $\tau$. This way,

$$\tau '=\tau _{0} =t'_{2} -t'_{1},$$

$$\tau =t_{2} -t_{1} \quad (1)$$

But according to the inverse Lorentz transformations we have

$$t_{2} -t_{1} ={(t'_{2} -t'_{1} )+{v\over c^{2} } (x'_{2} -x'_{1} )\over \sqrt{1-v^{2} /c^{2} } } \quad (2)$$

Substituting (1) into (2) and noting that the moving clock is always at the same point in the moving reference frame $S'$, that is,

$$x'_{1} =x'_{2} \quad (3)$$

We obtain

$$\tau ={\tau _{0} \over \sqrt{1-v^{2} /c^{2} } } ,\qquad (t_{0} =\tau ') \quad (4) $$

This formula means that the time interval measured by the fixed clocks is greater than the time interval measured by the single moving clock. This means that the moving clock lags behind the fixed ones, that is, it slows down.

The animation below depicts light clocks.

enter image description here

Sure, from the point of view of "stationary" observer $S$ beam of light moves by hypotenuse, though in "moving" frame $S'$ beam of light jumps up and down.

Good link: https://arxiv.org/abs/physics/0512013


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.