Situation: We consider two inertial systems $1$ and $2$ in standard configuration, i.e. system $2$ is moving into the direction $x_1$ with a speed of $v$. We only consider one spatial axis. An object in uniform motion is observed, from system $1$ and $2$.

System $1$ sees the object passing through a spatial distance of $s_1$ in a time difference of $t_1$ and concludes that it has a speed of $v_1 = s_1/t_1$.

System $2$ sees the object passing through a spatial distance of $s_2$ in a time difference of $t_2$ and concludes that it has a speed of $v_2 = s_2/t_2$.

Now in system $2$ spatial distances are Lorentz contracted and temporal distances are Lorentz dilatated: $\displaystyle s_2 = s_1/\gamma$ and $t_2 = \gamma t_1$ where $$\gamma = {1 \over \sqrt{1 - v^2/c^2}}$$ is the Lorentz factor. Thus:

$$v_2 = {s_2\over t_2} = {1\over \gamma^2} {s_1\over t_1} = \left(1 - {v^2 \over c^2}\right) v_1 \quad {\rm and\ thus}\quad v_1 = {v_2 \over 1 - v^2/c^2}$$

On the other hand, textbook velocity addition tells us that $$v_1 = v \oplus v_2 = {v + v_2 \over 1 + v\,v_2/c^2}$$ and, quite obviously, my derivation is wrong. This can be seen also from the fact that my "velocity formula" leads to an expression which is linear in $v_2$.

My question: Where does my analysis go wrong? Why do I get a different formula than the textbook velocity addition? I want to know what I am doing wrong here.

Note: I do understand that my analysis (obviously) is wrong and I am aware of the textbook derivation of velocity addition. I am interested in the exact point where my reasoning is wrong.

  • 2
    $\begingroup$ You need to re-learn (or properly learn) what Lorentz contraction and time dilation is actually saying. Those formulas cannot be applied here. $\endgroup$
    – Prahar
    Commented Apr 10, 2022 at 16:49
  • $\begingroup$ Just to add: time dilation and length contraction are themselves derived quantities, they are not suitable to use as a starting point because to define them in the first place you need to use the Lorentz Transform! $\endgroup$
    – m4r35n357
    Commented Apr 10, 2022 at 17:33
  • $\begingroup$ I added a note since it looks like the focus of my question was not clear enough. $\endgroup$ Commented Apr 10, 2022 at 20:53
  • 1
    $\begingroup$ Your reasoning is wrong as you substituted s2 by s1/gamma while this transformation is not valid. $\endgroup$ Commented Apr 10, 2022 at 21:05
  • $\begingroup$ Actually, this should be the formula for length contraction. It does not apply here because...?!?...which...requirement...?!? is not met? $\endgroup$ Commented Apr 10, 2022 at 21:11

2 Answers 2


Spatial distances are not contracted and times as well. The transformation rule is: $$x_1=\gamma (x_2+vt_2)$$ $$t_1=\gamma(t_2+vx_2/c^2)$$

A velocity $v_2$ measured by the second frame system will be measured by the first frame as: $$\frac{x_1}{t_1}=\frac{x_2+vt_2}{t_2+vx_2/c^2}=\frac{v_2t_2+vt_2}{t_2+vv_2t_2/c^2}=\frac{v_2+v}{1+vv_2/c^2}$$ supposing that the particle with $v_2$ start at the origin and both systems' origin coincide. These calculations have been carried out knowing that $x_2=v_2t_2$ and then dividing the numerator and denominator by $t_2$.

The first two equations can give the length contraction and time dilation. if you think about it Look: $$x_1=\gamma (x_2)$$ if $t_2$ is zero, and $$t_1=\gamma(t_2)$$ if $x_2$ is zero. But, why does $x_1$ depend on both $x_2$ and $t_2$? In special relativity, time is also a measurable coordinate as space. Indeed if you move respect a system $S_1$, you will not "see" just everything contracted, you will "see" the future of $S_1$ in front of you and the past of $S_1$ behind you. That means, your present will be his future and past. The implications of this have many important consequences, like the fact that what an observer $S_1$ measures as simultaneous (like two bombs at different places exploding at t=0) you may not (you might measure one explosion at t=1 and another at t=4). Minkowski diagrams will help you see it a lot better. So, that means that an interval of space for $S_1$, that means $\Delta x_1\neq 0$ and $\Delta t_1 =0$ may be an interval of space $\Delta x_2 \neq 0$ and also an interval of time $\Delta t_2 \neq 0$ for $S_2$. I insist on the fact that Minkowski diagrams will help you much better.


(Ok. Comment editing time and length is limited. For the sake of completeness, I will put down what I see as the final answer thus far in a bit broader wording. May it help the next person passing by this place.)

When substituting $s_2$ by $s_1/\gamma$ I do not claim to "invoke a transformation" but I apply a proposition stating that a measured length of an object is shorter than its proper length by the Lorentz factor.

This is where I go wrong: $s_1$ is not a proper length in system 1 but just a (normal) length in system 1. The proposition does not apply here.

The reason of the problem being that - trained as a mathematician - I expect all preconditions of a proposition to be stated precisely in that proposition and not being dispersed somewhere in the explanation. All the texts I consulted on this actually did have that restriction somewhere in the same chapter but not as part of the "proposition". That is the reason I missed the limitation in the applicability of the formula.


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.