I don't understand how to derive the matrix representing the Lorentz-transformation given two systems S and S':

$$x' = \Lambda x$$

these transformations do not leave the differences $\Delta x^\mu$ unchanged, but multiply them also by the matrix $\Lambda$:

$$(\Delta s)^2 = (\Delta x)^T \eta (\Delta x) = (\Delta x')^T \eta (\Delta x') = (\Delta x)^T \Lambda^T \eta \Lambda (\Delta x) \tag{1.26}$$

and therefore

$$\eta = \Lambda^T \eta \Lambda$$

I don't understand the mathematical passages in eq. 1.26 particularly:

  1. why is it needed to multiply by $\Lambda$?
  2. The role of the transpose symbol

I know that the space-time interval is given by

$$(\Delta s)^2 = \eta_{\mu \nu}\Delta x^\mu\Delta x^\nu$$

and I understand that the metric given by $\eta$ should be the same in every reference system.

  • 1
    $\begingroup$ I am not sure I understand as to which step you don't understand. $\endgroup$
    – ACat
    May 18, 2019 at 10:19

2 Answers 2


The role of the transpose symbol

The vectors are usually represented by the column matrix. $$\Delta x= \begin{bmatrix}c\Delta t \\ \Delta x \\\Delta y\\ \Delta z \end{bmatrix} .$$ Thus $$\Delta x^T= \begin{bmatrix}c\Delta t & \Delta x &\Delta y& \Delta z \end{bmatrix} .$$ The invariant interval in matrix representation is given by $$\Delta s^2=\Delta x^T\eta\Delta x=\begin{bmatrix}c\Delta t & \Delta x &\Delta y& \Delta z \end{bmatrix}\begin{bmatrix} 1& 0 & 0 & 0\\ 0 & -1 & 0 & 0 \\ 0 & 0 & -1 & 0 \\ 0& 0 & 0 & -1 \end{bmatrix}\begin{bmatrix}c\Delta t \\ \Delta x \\\Delta y\\ \Delta z \end{bmatrix}$$ $$=(c\Delta t)^2 - \Delta x^2 -\Delta y^2 - \Delta z^2$$

Why is it needed to multiply by Λ?

Because the Lorentz transformation equation is given by $$\Delta x' = \Lambda \Delta x$$ substituting this in the interval $(\Delta s)^2 = (\Delta x')^T \eta (\Delta x')$ we get $$ (\Delta s)^2= (\Lambda\Delta x)^T \eta (\Lambda\Delta x)=\Delta x^T\Lambda^T\eta \Lambda\Delta x$$
This is same as $(\Delta x)^T \eta (\Delta x)$ . Analyse it with the above equation, you will obtain $\eta = \Lambda^T \eta \Lambda$


To answer your question:

  1. If you look at $\mathrm{(1.26)}$ carefully, you will see that we have $$(\Delta s)^2 = (\Delta x )^T \eta(\Delta x) = (\Delta x')^T \eta(\Delta x')$$ This is equating the line element in the two different inertial frames. Notice the expression on the right hand side is a function of $\Delta x'$ not $\Delta x$. Using $\Delta x' = \Lambda \Delta x$ and $(\Delta x')^T = (\Delta x)^T \Lambda^T$ you can see where the both $\Lambda$ and $\Lambda^T$ have come from.

  2. The transpose symbol is because the line element must be a scalar, so the "matrix" that represents both $\Delta x$ terms must have inverse dimensions to the "matrix" that represents the metric. I personally don't like this formalism, and if you go on to do general relativity you will be introduced to tensors and the Einstein summation convention. In this formalism it is a lot clearer because the line element is given as $$ds^2=g_{\mu \nu} dx^\mu dx^\nu$$ where $g_{\mu \nu}$ is the metric. In this notation you can clearly see that the line element is a scalar and that metric is a (0,2) tensor so $dx^\mu dx^\nu$, if treated like a single object like it is in your question, must be a rank (2, 0) tensor.

I must admit I personally do not like the form given in $\mathrm{(1.26)}$ since I feel it is confusing as to what $\Delta x$ is and why, as you asked, we want to take the transpose. So I will write the same derivation in the notation used in general relativity, hopefully you may find this clearer.

We have the line element in the two frames as $$ds^2 = \eta_{\mu \nu}dx^\mu dx^\nu = \eta'_{\mu' \nu'} dx^{\mu'} dx^{\nu'}$$ where we are working in Minkowski space (special relativity) so $\eta = \eta'$. The $\mu'$ and $\nu'$ are to denote that these are new indices in the new frame. These are given by the Lorentz transformation $$dx^{\mu'} = \Lambda^{\mu'}{}_\mu dx^\mu.$$ Notice that here the transmormation "matrix" is actually a (1,1) tensor. Einstein summation convention tells us we sum over the repeated $\mu$ so what the only free index on the right-hand side is $\mu'$, just like on the left-hand side. Substituting this into the above gives $$\eta_{\mu \nu}dx^\mu dx^\nu = \eta'_{\mu' \nu'} \Lambda^{\mu'}{}_\mu \Lambda^{\nu '}{}_\nu dx^\mu dx^\nu$$ So, $$\eta_{\mu \nu} = \Lambda^{\mu'}{}_\mu \Lambda^{\nu'}{}_{\nu} \eta'_{\mu' \nu'}$$ Which is the same as $$\eta'=\Lambda^T\eta\Lambda.$$


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.