I'm attempting the problem shown below. Using the hint, I have so far found:

\begin{align}B^T \eta B &= \begin{pmatrix}\gamma & -\gamma\beta^j \\ -\gamma \beta_k & \delta_k^j+\frac{(\gamma -1)\beta^j \beta_k}{\beta^2}\end{pmatrix}. \begin{pmatrix}-1 & 0 \\ 0 & 1\end{pmatrix}. \begin{pmatrix}\gamma & -\gamma \beta_n\\ -\gamma\beta^m & \delta_n^m+\frac{(\gamma -1)\beta^m \beta_n}{\beta^2}\end{pmatrix}\\&= \begin{pmatrix}-\gamma^2(1-\beta^j \beta^m) & \gamma^2 \beta_n-\gamma \beta^j\Big(\delta_n^m + \frac{(\gamma -1)\beta^m \beta_n}{\beta ^2}\Big) \\ \gamma^2 \beta_k -\gamma \beta^m \Big(\delta_k^j + \frac{(\gamma -1)\beta^j \beta_k}{\beta^2}\Big) & -\gamma^2 \beta_k \beta_n + \Big(\delta_k^j + \frac{(\gamma -1)\beta^j \beta_k}{\beta^2}\Big)\Big(\delta_n^m + \frac{(\gamma -1)\beta^m \beta_n}{\beta^2}\Big)\end{pmatrix}\end{align}

However, I don't understand what is meant by time-time, time-space, space-time and space-space components. Some elaboration on what this means would be appreciated. Am I supposed to equate the above matrix components with $$\begin{pmatrix}-1 & 0 \\ 0 & 1\end{pmatrix}.$$

If so, what am I trying to solve for to truly show this is in the Lorentz group?

Please note, I'm only a novice with the summation notation, so I apologize if I've written anything out incorrectly.

Problem I'm Attempting



2 Answers 2


Hints :

$$ \text{your matrix :}\:\: \begin{pmatrix} -1 & 0 \\ 0 & 1 \end{pmatrix} \text{is this matrix :}\:\: \eta= \begin{bmatrix} -1 & 0 & 0 & 0\\ 0 & 1 & 0 & 0\\ 0 & 0 & 1 & 0\\ 0 & 0 & 0 & 1 \end{bmatrix} = \begin{bmatrix} -1 & \boldsymbol{0}^{\mathsf{T}} \\ \boldsymbol{0} & \mathrm{I} \end{bmatrix} \tag{01} $$ where $$ \boldsymbol{0}\equiv \begin{bmatrix} 0\\ 0\\ 0 \end{bmatrix} =\text{the null column vector}, \qquad \boldsymbol{0}^{\mathsf{T}}\equiv \begin{bmatrix} 0 & 0 & 0 \end{bmatrix} =\text{the null row vector} \tag{02} $$

$$ B= \begin{bmatrix} \gamma & -\gamma\beta^j \\ -\gamma \beta_k & \delta_k^j+\dfrac{(\gamma -1)\beta^j \beta_k}{\beta^2} \end{bmatrix} = \begin{bmatrix} \gamma & -\gamma\boldsymbol{\beta}^{\mathsf{T}} \\ -\gamma \boldsymbol{\beta} & \mathrm{I}+\dfrac{(\gamma -1)\boldsymbol{\beta} \boldsymbol{\beta}^{\mathsf{T}}}{\beta^2} \end{bmatrix} \tag{03} $$

$$ \boldsymbol{\beta}= \begin{bmatrix} \beta_{1}\\ \beta_{2}\\ \beta_{3} \end{bmatrix}, \qquad \boldsymbol{\beta}^{\mathsf{T}}= \begin{bmatrix} \beta^{1} & \beta^{2} & \beta^{3} \end{bmatrix}, \qquad \mathrm{I}= \begin{bmatrix} 1&0&0\\ 0&1&0\\ 0&0&1 \end{bmatrix} \tag{04} $$

$$ \boldsymbol{\beta}^{\mathsf{T}}\boldsymbol{\beta}= \begin{bmatrix} \beta^{1} & \beta^{2} & \beta^{3} \end{bmatrix} \begin{bmatrix} \beta_{1}\\ \beta_{2}\\ \beta_{3} \end{bmatrix} =\beta^{m}\beta_{m} =\boldsymbol{\beta}\boldsymbol{\cdot}\boldsymbol{\beta}=\Vert \boldsymbol{\beta} \Vert ^{2}=\beta^{2}=\dfrac{v^2}{c^2}=\dfrac{\gamma^2-1}{\gamma^2} \tag{05} $$

$$ \boldsymbol{\beta}\boldsymbol{\beta}^{\mathsf{T}}= \begin{bmatrix} \beta^{1}\\ \beta^{2}\\ \beta^{3} \end{bmatrix} \begin{bmatrix} \beta_{1} & \beta_{2} & \beta_{3} \end{bmatrix} =\beta_{m}\beta^{n}= \begin{bmatrix} \beta_{1}^{2} & \beta_{1}\beta_{2} & \beta_{1}\beta_{3}\\ \beta_{2}\beta_{1} & \beta_{2}^{2} & \beta_{2}\beta_{3}\\ \beta_{3}\beta_{1} & \beta_{3}\beta_{2} & \beta_{3}^{2} \end{bmatrix} \tag{06} $$

$$ \left(\boldsymbol{\beta}\boldsymbol{\beta}^{\mathsf{T}}\right)^{2}= \left(\boldsymbol{\beta}\boldsymbol{\beta}^{\mathsf{T}}\right)\left(\boldsymbol{\beta}\boldsymbol{\beta}^{\mathsf{T}}\right)= \boldsymbol{\beta}\underbrace{ \left (\boldsymbol{\beta}^{\mathsf{T}} \boldsymbol{\beta}\right)}_{\Vert \boldsymbol{\beta} \Vert ^{2}}\boldsymbol{\beta}^{\mathsf{T}}= \Vert \boldsymbol{\beta} \Vert ^{2}\boldsymbol{\beta}\boldsymbol{\beta}^{\mathsf{T}}=\left(\dfrac{{\gamma}^2-1}{\gamma^2}\right)\boldsymbol{\beta}\boldsymbol{\beta}^{\mathsf{T}} \tag{07} $$

Note that if $\:\mathbf{n}\:$ is a unit 3-vector then $\:\mathrm{P}_{\mathbf{n}}=\mathbf{n}\mathbf{n}^{\mathsf{T}}\:$ is the projection on its direction, since for every $\:\mathbf{x} \in \mathbb{R}^{3}\:$ $$ \mathrm{P}_{\mathbf{n}}\mathbf{x}=\mathbf{n}\mathbf{n}^{\mathsf{T}} \mathbf{x}= \begin{bmatrix} n_{1}\\ n_{2}\\ n_{3} \end{bmatrix} \underbrace{ \begin{bmatrix} n_{1} & n_{2} &n_{3} \end{bmatrix} \begin{bmatrix} x_{1}\\ x_{2}\\ x_{3} \end{bmatrix}}_{\left(\mathbf{x} \boldsymbol{\cdot}\mathbf{n} \right)} =\left(\mathbf{n} \boldsymbol{\cdot}\mathbf{x} \right)\mathbf{n} \tag{08} $$ with the well known property of projections $$ \mathrm{P}_{\mathbf{n}}^{2}=\mathrm{P}_{\mathbf{n}} \tag{09} $$ Defining $$ \mathbf{n} \equiv \dfrac{\boldsymbol{\beta}}{\Vert \boldsymbol{\beta} \Vert}=\dfrac{\boldsymbol{\beta}}{\beta} \tag{10} $$ then $$ \mathrm{P}_{\mathbf{n}}=\mathbf{n}\mathbf{n}^{\mathsf{T}}=\left( \dfrac{ \boldsymbol{\beta}}{\beta} \right) \left(\dfrac{\boldsymbol{\beta}}{\beta}\right)^{\mathsf{T}}=\dfrac{\boldsymbol{\beta}\boldsymbol{\beta}^{\mathsf{T}} }{\beta^{2}} \tag{11} $$ and (07) is its property as projection.


\begin{align} & B^{\mathsf{T}}\eta B =\\ &\begin{bmatrix} \gamma & -\gamma\boldsymbol{\beta}^{\mathsf{T}} \\ -\gamma \boldsymbol{\beta} & \mathrm{I}+\dfrac{(\gamma -1)\boldsymbol{\beta} \boldsymbol{\beta}^{\mathsf{T}}}{\beta^2} \end{bmatrix} \begin{bmatrix} -1 & \boldsymbol{0}^{\mathsf{T}}\\ &\\ \boldsymbol{0} & \mathrm{I} \end{bmatrix} \begin{bmatrix} \gamma & -\gamma\boldsymbol{\beta}^{\mathsf{T}} \\ -\gamma \boldsymbol{\beta} & \mathrm{I}+\dfrac{(\gamma -1)\boldsymbol{\beta} \boldsymbol{\beta}^{\mathsf{T}}}{\beta^2} \end{bmatrix}=\\ &\begin{bmatrix} \gamma & -\gamma\boldsymbol{\beta}^{\mathsf{T}} \\ -\gamma \boldsymbol{\beta} & \mathrm{I}+\dfrac{(\gamma -1)\boldsymbol{\beta} \boldsymbol{\beta}^{\mathsf{T}}}{\beta^2} \end{bmatrix} \begin{bmatrix} -\gamma & +\gamma\boldsymbol{\beta}^{\mathsf{T}} \\ -\gamma \boldsymbol{\beta} & \mathrm{I}+\dfrac{(\gamma -1)\boldsymbol{\beta} \boldsymbol{\beta}^{\mathsf{T}}}{\beta^2} \end{bmatrix}= \begin{bmatrix} \sigma & \boldsymbol{\rho}^{\mathsf{T}} \\ &\\ \boldsymbol{\rho} & \mathrm{Z} \end{bmatrix}\equiv \xi \tag{12} \end{align}

where $\:\xi\:$ a real symmetric $\:4\times 4\:$ matrix with elements $\:\sigma,\boldsymbol{\rho},\mathrm{Z}\:$ a real scalar, a real 3-vector and a real symmetric $\:3\times 3\:$ matrix respectively, all to be determined. Now, \begin{align} \sigma=-\gamma^{2}\underbrace{\left(1-\boldsymbol{\beta}^{\mathsf{T}}\boldsymbol{\beta}\right)}_{1-\tfrac{v^2}{c^2}=\gamma^{-2}} \quad \Longrightarrow \quad \sigma=-1 \tag{13a} \end{align}

\begin{align} \boldsymbol{\rho} & =\gamma^{2}\boldsymbol{\beta}-\gamma\boldsymbol{\beta}-\dfrac{\gamma(\gamma -1)\boldsymbol{\beta}\boldsymbol{\beta}^{\mathsf{T}} \boldsymbol{\beta}}{\beta^2}\\ & =\gamma\left(\gamma-1\right)\boldsymbol{\beta}-\gamma\left(\gamma-1\right)\boldsymbol{\beta}\underbrace{\left(\dfrac{\boldsymbol{\beta}^{\mathsf{T}}\boldsymbol{\beta}}{\beta^2}\right)}_{=1} \quad \Longrightarrow \quad \boldsymbol{\rho}=\boldsymbol{0} \tag{13b} \end{align}

\begin{align} \mathrm{Z} & =-\gamma^{2}\boldsymbol{\beta}\boldsymbol{\beta}^{\mathsf{T}}+\biggl(\mathrm{I}+\dfrac{(\gamma -1)\boldsymbol{\beta} \boldsymbol{\beta}^{\mathsf{T}}}{\beta^2}\biggr)^{2}\\ & = -\gamma^{2}\boldsymbol{\beta}\boldsymbol{\beta}^{\mathsf{T}}+\mathrm{I}+\dfrac{2(\gamma -1)\boldsymbol{\beta} \boldsymbol{\beta}^{\mathsf{T}}}{\beta^2}+\left(\gamma -1\right)^{2}\overbrace{\biggl(\dfrac{\boldsymbol{\beta} \boldsymbol{\beta}^{\mathsf{T}}}{\beta^2}\biggr)^{2}}^{=\boldsymbol{\beta}\boldsymbol{\beta}^{\mathsf{T}}/\beta^2}\\ & = \mathrm{I}+ \underbrace{\left[-\gamma^{2}+\dfrac{2(\gamma -1)}{\beta^2}+\dfrac{(\gamma -1)^{2}}{\beta^2} \right]}_{=0}\boldsymbol{\beta}\boldsymbol{\beta}^{\mathsf{T}}\quad \Longrightarrow \quad \mathrm{Z}=\mathrm{I} \tag{13c} \end{align}

So, \begin{equation} B^{\mathsf{T}}\eta B = \xi = \begin{bmatrix} \sigma & \boldsymbol{\rho}^{\mathsf{T}} \\ &\\ \boldsymbol{\rho} & \mathrm{Z} \end{bmatrix}= \begin{bmatrix} -1 & \boldsymbol{0}^{\mathsf{T}}\\ &\\ \boldsymbol{0} & \mathrm{I} \end{bmatrix} \equiv \eta \tag{14} \end{equation}


  • $\begingroup$ Thanks. I ended up solving the problem, although I did stick with the summation notation. Your hints did help. $\endgroup$
    – Akyidrian
    Jun 7, 2016 at 6:21
  • $\begingroup$ @Akyidrian : As a rule of this site we must not post complete answers to homework-like exercises but hints only. Since you did solve the problem by yoursefl using summation notation, I complete my answer with vectors. I think you'll find it useful for comparison with your solution and for future use. $\endgroup$
    – Frobenius
    Jun 7, 2016 at 8:12

Another, less messy way to do this is as follows. Link the said matrix to the identity matrix by a path defined by:

$$\Lambda:\mathbb{R}\to \mathscr{M}_{4\times4};\;\Lambda(\zeta) = \left(\begin{array}{c|c}\cosh\zeta & -\hat{B}^T \,\sinh\zeta\\\hline -\hat{B}\,\sinh\zeta &\mathrm{id} +\hat{B}\,\hat{B}^T\, (\cosh\zeta-1) \end{array}\right)\tag{1}$$

where $\zeta = \operatorname{artanh}\frac{v}{c}$ is the rapidity of the putative boost. Here $\hat{B} =\frac{1}{v} \begin{pmatrix}v_1\\v_2\\v_3\end{pmatrix}$ is the unit vector of direction cosines pointing along the direction of the boost.

Exercise:Check that all matrices of the stated form can be written in the form of (1), so that they all form a smooth path through the identity (where they pass when $\zeta=0$).

Note the useful little formula $\hat{B}^T\,\hat{B} = 1$, so that you can almost manipulate matrices of the form in (1) as though their elements were scalar, aside from that $\hat{B}\,\hat{B}^T$ is left unsimplified. You get things like $(\hat{B}\,\hat{B}^T)^N = \hat{B}\,\hat{B}^T;\,N\geq 1$ (this means that $(\hat{B}\,\hat{B}^T)^N$ is the idempotent projector onto the direction of the boost) which you can use in the following.

Excercise: Prove that

$$\Lambda(\zeta)\Lambda(-\zeta) = \mathrm{id}\tag{2}$$


$$\frac{\mathrm{d}\Lambda}{\mathrm{d}\zeta} \Lambda^{-1} = -\left(\begin{array}{c|c}0&\hat{B}^T\\\hline\hat{B}&0\end{array}\right)\tag{3}$$

Now, given (3), we have very simply:

$$\frac{\mathrm{d}}{\mathrm{d}\zeta} \left(\Lambda^T\,\eta\,\Lambda\right) = \Lambda^T\left(-\left(\begin{array}{c|c}0&\hat{B}^T\\\hline\hat{B}&0\end{array}\right)^T\,\eta - \eta \,\left(\begin{array}{c|c}0&\hat{B}^T\\\hline\hat{B}&0\end{array}\right)\right)\,\Lambda=0\tag{4}$$

and, since the sought identity follows trivially for $\zeta=0$, we have, through (4) a Cauchy initial value problem wherein the the derivative is a Lipschitz-continuous function of $\Lambda^T\,\eta\,\Lambda$, therefore $\Lambda^T\,\eta\,\Lambda=\eta$, true for all $\zeta$, is the unique solution to this CIVP and the identity is proven.

You've therefore got yourself a rather useful and compact expression for a general boost in (1), and you can see, in the light of (3), that it is the matrix exponential of $\zeta$ times the simple matrix on the right hand side of (3). The matrix on the right hand side of (3) is sometimes called an infinitessimal boost; all infinitessimal boosts are linear combinations, with the direction cosines as weights, of the three infinitessimal boosts for the three spatial co-ordinate directions.


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.