I am looking for references on contact structures in the framework of pseudo-riemannian manifolds. For instance on the Lorentz-Minkowski $3$-space (resp. $(n+1)$-space). Denoting by $L$ (resp. $H$) the Lorentz-Minkowski $3$-space (resp. the hyperbolic plane regarded as pseudo-sphere in $L$), it seems to me that $w= udx+vdy-wdz, (x,y,z,u,v,w)$ in $LxH$, defines a contact structure on $LxH$. It is as the classical example of the unit tangent bundle of $3$-dimensional Euclidean space E but replacing $E$ by $L$ and the Euclidean scalar product by the Lorentzian one in $\left<(dx,dy,dz),(u,v,w)\right>$.

  • $\begingroup$ I once heard an interesting talk by nemirovwski about such things. This is probably contained in this paper: arxiv.org/pdf/0810.5091.pdf, but I have not read this myself $\endgroup$
    – Thomas Rot
    Commented Jul 11, 2017 at 19:15

1 Answer 1


Yes, this is a contact structure. The point is that, using the pseudo-Riemannian metric, you can identify its future-directed 'unit' tangent bundle with an open set in the projectivized cotangent bundle. Since the projectivized cotangent bundle of any manifold carries a canonical contact structure, this identification induces a canonical contact structure on the future-directed 'unit' tangent bundle. In your case, $L\times H$ is exactly the future-directed unit tangent bundle. As you note, it corresponds to the unit sphere bundle of a Riemannian manifold (which also inherits a contact structure in exactly the same way).


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.