In mechanics energy is $E = \frac{m v^2}{2}$

The corresponding relativistic equation is $ E = m (\gamma -1) c^2 $ which for v<<c is appoximately $\frac{m v^2}{2}$

I know that the equation above is correct because I have seen the derivation at Wikipedia.

But energy can also be calculated by $E = f d$

The corresponding relativistic equation would be four-force times displacement four-vector (i.e. four-position)

$ E = \left(\gamma {\mathbf{f}\cdot\mathbf{v} \over c},\gamma{\mathbf f}\right) \cdot \left(ct, \mathbf{r}\right) $

Is there a way to show that this second relativistic equation gives a value for energy that doesnt contradict the first equation above?

($ct$ has units of distance. $\frac{v}{c}$ is dimensionless and so is $\gamma$)

f is the rate of change of proper momentum (mass times proper velocity)

${\mathbf f}={\mathrm{d} \over \mathrm{d}t} \left(\gamma m {\mathbf v} \right)={\mathrm{d}\mathbf{p} \over \mathrm{d}t}$


${\mathbf{f}\cdot\mathbf{v}}={\mathrm{d} \over \mathrm{d}t} \left(\gamma mc^2 \right) $

The derivative of gamma is:

$\dot\gamma = \frac{d \gamma}{d t} = \frac{d \gamma}{dv} \frac{dv}{dt} = \frac{v \gamma^3 a}{c^2}$

  • $\begingroup$ There is no term "proper velocity". If it would exist then it would be the zero 3-vector. If you mean proper 4-velocity vector then it would be $\mathbf U_0 \boldsymbol{=}(c, \boldsymbol{0})$. $\endgroup$
    – Frobenius
    Dec 21, 2020 at 14:15
  • 2
    $\begingroup$ @Frobenius sometimes “celerity” is called “proper velocity”: en.m.wikipedia.org/wiki/Proper_velocity But I think “proper velocity” is terrible terminology and would always use the term “celerity” instead. $\endgroup$
    – Dale
    Dec 21, 2020 at 14:32
  • 1
    $\begingroup$ @Dale : Thanks for the info. $\endgroup$
    – Frobenius
    Dec 21, 2020 at 14:39

2 Answers 2


It's pretty easy to see that if you take the inner product of the force four-vector with a displacement four-vector, you don't get a correct expression for the mechanical work done by the force. This is because the inner product of two four-vectors is a scalar, which is the same in all reference frames. But energy obviously depends on your frame of reference. A more compact way of expressing this is that in terms of three-vectors, $\textbf{F}\cdot\textbf{v}$ is an expression for the power, while in terms of four-vectors, this expression vanishes identically.

It is true, however, that if you express work in terms of the force and displacement three-vectors, the result is relativistically valid, and you don't need to introduce factors of gamma or anything like that. There is a compact proof of this fact (here given in one dimension):

$$\frac{ d E}{ d x} = \frac{ d E}{ d p}\frac{ d p}{ d t} \frac{ d t}{ d x} = \frac{ d E}{ d p} \frac{F}{v}$$

The desired result follows from application of the identity $dE/dp=v$.

For a more detailed discussion of this kind of thing, see ch. 4 of my SR book, http://lightandmatter.com/sr/ .


To reconcile the 4D calculation, you need $(dE/dt,dp/dt)\cdot(dt,dx)$ for infinitesimal 4-displacement

Take the units where $c=1$. It's better to regard $E=m\gamma$ instead of $E=m(\gamma-1)$ although it's just a constant shift. The Minkowski-norm $(E,p)\cdot (E,p)=E^2-p^2=m^2\gamma^2-m^2 v^2\gamma^2=m^2$ is a constant, the static mass squared as expected. So the differential of that is identically 0, which gives $0=2(dE/dt,dp/dt)\cdot(E,p)=2(E\;dE/dt-dp/dt\cdot p).$ So you arrive at $dE/dt=dp/dt \cdot (p/E)=dp/dt\cdot dx/dt,$ which is consistent with $(dE/dt,dp/dt)\cdot(dt,dx)=0$.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.