0
$\begingroup$

Theorem: Suppose $A_{\mu}$ is a $4$-vector and $B^{\mu}$ is an object with $4$ components. If $A_{\mu}B^{\mu}$ is a scalar then $B^{\mu}$ is a $4$-vector.

I have been stuck on trying to prove this theorem for quite a while (See Ref.1) and haven't made much progress. I think this is due to my lack of understanding on what it means to be a $4$-vector.

What I have tried so far is to use the invariant quantities $A_{\mu}A^{\mu}$ and $A_{\mu}B^{\mu}$ to show $B_{\mu}B^{\mu}$ is an invariant quantity.

I would like to know if this approach makes sense, since the only quantifiable property of a $4$-vector I can think of is its Minkowski inner product with itself being invariant.

I am new to this subject and would be really grateful for any help or insights.

References:

  1. L. Susskind & A. Friedman, Special Relativity and Classical Field Theory: The Theoretical Minimum , 2017; p. 181 section 5.5 .
$\endgroup$
2
  • 1
    $\begingroup$ This is a special case of so-called quotient theorem in tensor analysis. Generally, it doesn't matter what dimension of space is, but simply a result that concerns the order of tensors. $\endgroup$
    – GK1202
    Commented Jun 6, 2020 at 18:35
  • 1
    $\begingroup$ THEOREM : Suppose $A_\mu$ is an A R B I T R A R Y 4-vector and $B^\mu$ is an object with 4 components. If $A_{\mu} B^\mu$ is invariant then $B^\mu$ is a 4-vector. $\endgroup$
    – Voulkos
    Commented Jun 6, 2020 at 20:06

1 Answer 1

1
$\begingroup$

From $^{\prime\prime}$Tensor Calculus$^{\prime\prime}$ by J.L.Synge-A.Schild, Dover Edition 1978 :

1.6. Tests for tensor character. The direct test for the tensor character of a set of quantities is this: see whether the components obey the law of tensor transformation when the coordinates are changed. However, it is sometimes much more convenient to proceed indirectly as follows.

Suppose that $\,A_{r}\,$ is a set of quantities which we wish to test for tensor character. Let $\,X^{r}\,$ be the components of an arbitrary contravariant tensor of the first order. We shall now prove that if the inner product $\,A_{r}X^{r}\,$ is an invariant, then $\,A_{r}\,$ are the components of a covariant tensor of the first order. We have, by the given invariance,

\begin{equation} A_{r}X^{r}\boldsymbol{=}A^{\prime}_{r}X^{\prime\, r} \tag{1.601.}\label{1.601.} \end{equation} and, by the law of tensor transformation, \begin{equation} X^{\prime\, r}\boldsymbol{=}X^{s}\dfrac{\partial x^{\prime\, r}}{\partial x^{s}} \tag{1.602.}\label{1.602.} \end{equation} Substituting this in the right-hand side of \eqref{1.601.}, rearranging, and making a simple change in notation, we have \begin{equation} \left(A_{s}\boldsymbol{-}A^{\prime}_{r}\dfrac{\partial x^{\prime\, r}}{\partial x^{s}}\right)X^{s}\boldsymbol{=}0 \tag{1.603.}\label{1.603.} \end{equation} Since the quantities $\,X^{s}\,$ are arbitrary, the quantity inside the parentheses vanishes; this establishes the tensor character of $\,A_{r}\,$, by \eqref{1.402.}

$\boldsymbol{=\!=\!=\!==\!=\!=\!==\!=\!=\!==\!=\!=\!==\!=\!=\!==\!=\!=\!==\!=\!=\!==\!=\!=\!==\!=\!=\!==\!=\!=\!==\!=\!=\!==\!=\!=\!=}$

A set of quantities $\,T_{r}\,$ are said to be the components of a covariant vector if they transform according to the equation \begin{equation} T_{r}\boldsymbol{=}T_{s}\dfrac{\partial x^{s}}{\partial x^{\prime\, r}} \tag{1.402.}\label{1.402.} \end{equation}

$\endgroup$
2
  • 2
    $\begingroup$ This is a misunderstanding. You do really need the scalar property only for a single arbitrary contravariant 4-vector as implied in the question. Even the passage you quote doesn't posit at the outset that it has to be true "for all" such vectors, even though the proof then incorrectly assumes this stronger statement. You're partly correct in that the theorem in the question isn't true when the arbitrary vector is an operator (because it leads to a differential equation with integration constants), but it is in fact true otherwise. $\endgroup$
    – Jens
    Commented Aug 19, 2021 at 0:55
  • $\begingroup$ @Jens : May be you are right. I'll check it, but I'll disagree with you only in case I could build a counterexample. $\endgroup$
    – Voulkos
    Commented Aug 19, 2021 at 5:44

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.