I was wondering about the idea of time and its measurement, i.e. clocks. As I understand, we define time by the count of a recurring event. Let's assume that we use two atomic clocks A and B to measure time, based on different atoms.

Now, assume the following scenario:

When clock A reads 1 tick, B reads 1 tick
When clock A reads 2 ticks, B reads 3 ticks
When clock A reads 3 ticks, B reads 4 ticks
When clock A reads 4 ticks, B reads 6 ticks
When clock A reads 5 ticks, B reads 7 ticks

And so on...

How do you pick a preferred clock here? And does the choice of clock define the laws of physics? A body in uniform motion as seen with clock A would not be in uniform motion as per clock B.

In general, the atomic clock situation I described above can be replaced with any natural recurring event, so how do we treat such asynchronous events? Or do we simply not find them in nature?

  • 2
    $\begingroup$ Why not? This is exactly the soul of the question. Isn't it possible for two recurring events in nature to not be in agreement with one-another? $\endgroup$ Mar 21, 2020 at 23:17
  • 2
    $\begingroup$ I expect the only answer is that we happen to find a whole lot of different clocks that tick in a way commensurate with A and very few that are commensurate with B. But I'll be very interested to see if someone has a better answer. $\endgroup$
    – WillO
    Mar 21, 2020 at 23:25
  • 1
    $\begingroup$ We use time as the evolution parameter for many physical phenomena (and we all agree to use clocks based on something called "seconds" but that's all it is, an agreement), but we could use another parameter to describe the same phenomena. "Laws of physics" would change in the sense we would need to re-state its axioms based on a new definition of "time", but the physics behind them would be the same. Nature doesn't care how we describe it. $\endgroup$
    – peguerosdc
    Mar 22, 2020 at 0:00
  • 2
    $\begingroup$ Clocks should be periodic. Clock B isn't a clock. $\endgroup$
    – d_b
    Mar 22, 2020 at 17:29
  • 3
    $\begingroup$ @d_b it could be clock A that is not periodic. The point of the question is that given only this data it is not clear which clock is non periodic. $\endgroup$
    – Dale
    Mar 22, 2020 at 18:17

3 Answers 3


How do you pick a preferred clock here?

You get another clock. This is essentially the reason that Hafele–Keating used an ensemble of four clocks in their famous experiment. The average of four clocks is more accurate than a single clock and if one clock has a malfunction then it will be apparent which clock is problematic. Preferred clocks are ones that consistently agree with each other the best.

  • $\begingroup$ What do you mean get another clock? It's not that any clock is malfunctioning, it's just the way it is $\endgroup$ Mar 22, 2020 at 18:21
  • $\begingroup$ I mean literally get (purchase or build) a third clock. If the third clock agrees with A then you prefer A. If the third clock agrees with B then you prefer B. If you are still not confident with just three clocks then get a fourth or fifth or ... $\endgroup$
    – Dale
    Mar 22, 2020 at 18:38
  • $\begingroup$ OK, it seems that in your line of reasoning, you've an underlying assumption, that there is indeed one true clock that we are trying to find and most clocks should conform to that 'true' clock. Do I get it right? But my question persists, each clock is based on an independent phenomenon, why should we expect them to agree? $\endgroup$ Mar 22, 2020 at 18:51
  • 2
    $\begingroup$ No. The reasoning is the opposite. There is no one true clock. The reasoning is simply that good measuring devices agree with each other well, and even better measurement devices agree with each other even better. That isn’t specific to clocks but applies to all measurement devices. $\endgroup$
    – Dale
    Mar 22, 2020 at 18:59
  • 1
    $\begingroup$ Ok, I see how this sort of answers the question. But this is not exactly what I meant. Maybe I need to think more on it and come out with a more clear thought. Accepting it for now. $\endgroup$ Mar 22, 2020 at 19:06

To begin with, we must eliminate the possibility that one of the two clocks is broken. Once we are sure that is not the case, the laws of physics indeed depend on our choice of clock. It happens to be that in the case of our universe, we have principles that say that the laws of physics are the independent of the choice of reference frame. However, remember that we had to establish first that the speed of light was the same in all inertial frames, before we could postulate that there was no preferred inertial frame. This is certainly not an axiom we can just start with.

  • $\begingroup$ ok, clarification: until we have proved the second statement to be true, the first has to be assumed. in our universe, we have proved it, but ours is not the universe being described in the question. $\endgroup$
    – NewUser
    Mar 22, 2020 at 0:51
  • $\begingroup$ sure, i am indeed working with the assumption that neither is broken. that brings us to the question: is there a preferred clock? my argument is, there need not be. $\endgroup$
    – NewUser
    Mar 22, 2020 at 22:24
  • $\begingroup$ Oh, I agree! Sorry, I think I intended to make this comment elsewhere - not sure what happened! I'll delete it. $\endgroup$ Mar 23, 2020 at 0:58

At an in-principle level, no, it doesn't change the laws of physics, but it could change how we write the equations we use to describe them.

The laws of physics depend on the dimension Time. Any given clock is simply an instrument to measure time. The standard scientific unit, which we measure in, for time is the second.

The second is currently defined based on the vibration of certain atoms, but was originally defined as a fraction of an minute which was in turn defined as a fraction of an hour and so on - ultimately the second was originally based on the rotation of the earth. This could be considered one "clock" in the way you mean, if i understand you correctly.

Changing the definition to be based on atomic vibrations allowed a more accurate definition of the size of the unit, but didn't change any fundamental physics, which ultimately relate to the dimension, time. Choosing to define it based on the vibrations of a different atom wouldn't have changed anything except the ease/accuracy/reliability with which we could measure the unit.

If you mean choosing a different unit size, other than what we currently know as a second, that would either have resulted in different choices of other unit sizes, or different looking fundamental equations - but they would only have differed by the constants needed to make the units work.

Take Newton's Second Law. The fundamental law is that force is equal to the rate of change of momentum. Given constant mass, that results in the familiar $$ F = ma $$ This is given use of kg for mass and meters per second per second for acceleration. However, if we wanted to measure acceleration in miles per hour per hour we could. The equation would then be $$ F = Kma $$ Where K is a constant (approximately 8052.985) to make the units work. The fundamental physics is the same.

In this case, we pretty much define the unit of force (Newtons) so that this holds (i.e. so that the constant K = 1 and can be left out altogether)

In practice, it would probably have resulted in changes in some other units to make things nicer, rather than just shoving in a bunch of constants.


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.