A closed achronal surface is Cauchy if its domain of dependence is entire manifold. From information given on a Cauchy surface, we can predict what happens throughout all of spacetime.

How do we determine whether a spacetime has a Cauchy surface or not? And how do we specify information on the Cauchy surface?

Also, given a Cauchy surface and the information on it, is it possible to predict whether the spacetime has a singularity or not?


There's a few handy ways to find out more or less easily if a hypersurface is a Cauchy surface.

Rather than checking every possible timelike curve, it is possible to simply check every inextendible null geodesics. By a theorem of Wald (theorem 8.3.7), for a closed, achronal, edgeless set $\Sigma$, $\Sigma$ is a Cauchy surface if those curves intersect $\Sigma$ and enter $I^+(\Sigma)$ and $I^-(\Sigma)$.

A useful tool for this is also the use of temporal functions. A temporal function $t$ is a function $t : \mathcal M \to \mathbb R$ that is smooth and with a gradient that's a past-directed timelike vector. The level sets of this functions (surfaces such that $t(\Sigma) = \text{const}$) are foliations of the manifold into spacelike hypersurfaces. The existence of a temporal function only guarantees that the spacetime is stably causal (Minkowski space minus a closed set would admit one, after all).

A globally hyperbolic spacetime also requires the splitting of the spacetime as a product $\mathbb R \times \Sigma$, with $g = -\beta^2 dt^2 \oplus h_t$, with $h_t$ a metric on $\Sigma$, $\beta$ a positive function and $t$ said temporal function ($dt$ is the $1$-form obtained from its exterior derivative). If you can obtain such a form for the spacetime it should be a good hint that the surfaces $t^{-1}(\text{const}) = \Sigma$ are Cauchy surfaces. As far as I know, the usual theorem goes the other way : the spacetime being globally hyperbolic guarantees the existence of this form. It is possible that a non-globally hyperbolic spacetime admits such a form, in particular I think anti-de Sitter space does. But that can be a useful tool combined with the null geodesics to find Cauchy surfaces.

And of course, use the invariance of causal structure under Weyl transform : if you have a simpler spacetime conformally related to yours, use that one to do your calculations


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.