In Ta-Pei Cheng's Relativity, Gravitation and Cosmology book, pg. 88, it was stated that the infinitesimal invariant interval $ds^2=g_{\mu\nu}dx^\mu dx^\nu$ for a spherically symmetric metric $g_{\mu\nu}$ is: $$ds^2=Ad\vec{r}\cdot d\vec{r}+B(\vec{r}\cdot d\vec{r})^2+Cdt(\vec{r}\cdot d\vec{r})+Ddt^2$$ where $A$, $B$, $C$ and $D$ are scalar functions of $t$ and $\vec{r}\cdot\vec{r}$.

Is it correct to say that for a general metric $g_{\mu\nu}$ (not necessarily spherically symmetric), the invariant interval $ds^2$ will be the similar as above, i.e.

$$ds^2=Ad\vec{r}\cdot d\vec{r}+B(\vec{r}\cdot d\vec{r})^2+Cdt(\vec{r}\cdot d\vec{r})+Ddt^2$$ but now with $A$, $B$, $C$ and $D$ being scalar functions of $t$ and $\vec{r}$?


3 Answers 3


To describe the most general spacetime metric you need 10 functions (the metric is a symmetric 4x4 matrix). However, using the transformation law for the metric, you can always choose a new coordinate system in which you eliminate 4 of the above 10 functions. Thus, the most general metric (regardless of whether you represent it in spherical coordinates) has 6 free functions. In your example you have only 4 functions: $A, B, C$ and $D$, so it cannot describe the most general spacetime.

By the way, note that the very meaning of the coordinate system is not given a priory. So, before you get to know more about the details of your spacetime, it is meaningless to say, for example, that $\vec{r}$ is the usual radial vector.


Not quite. There are more possible terms one can add that are forbidden by spherical symmetry. For example, rotations in some direction $\theta$ are characterised by a term $~dtd\theta$ in the metric. Expanding your metric using $d \vec{r} \cdot d \vec{r}=d r^{2}+r^{2}\left(d \theta^{2}+\sin ^{2} \theta d \phi^{2}\right)$, we see that

$$ d s^{2}=A\left[d r^{2}+r^{2}\left(d \theta^{2}+\sin ^{2} \theta d \phi^{2}\right)\right]+B r^{2} d r^{2}+C r d r d t+D d t^{2} $$

Since there are no terms that go like $dtd\theta$ or $dtd\phi$, we know that this can't be the most general metric.

  • $\begingroup$ Why the downvote? $\endgroup$
    – Akoben
    Nov 17, 2020 at 17:30

Spherical symmetry implies that the metric is independent of angles, that is, where you are standing on a spherical surface. Basically, entire surface of the sphere (at every r) looks the same everywhere on it. Mathematically, this translates to the metric coefficients, $A$, $B$, $C$ and $D$ being independent of $\theta$ and $\phi$ coordinates.

If you forgo spherical symmetry, in the most general metric, the coefficients $A$, $B$, $C$ and $D$ will be functions of $r$, $\theta$, and $\phi$.

About the dependence on $t$, if a metric is time-independent people generally add a qualifier of steady and/or stationary. So when you are talking about just a spherically symmetric metric, the metric coefficients depend on $t$ anyway in general.


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.