Lorentz manifolds modelled on a Lorentz symmetric space (Q1176459): Difference between revisions

Let \((M,g)\) be a pseudo-Riemannian manifold and \((M_ 0,g_ 0)\) a pseudo-Riemannian symmetric space of the same dimension and signature. One says that \((M,g)\) is modelled on \((M_ 0,g_ 0)\) if, for any point \(\xi\in M\), there exists an isometry \(\phi_ \xi: T_ \xi M\rightarrow T_{u_ 0}M_ 0\) (\(u_ 0\) being a base point of \(M_ 0\)) such that \(\varphi^*_ \xi R_{0,u_ 0}=R_ \xi\). This means that \((M,g)\) has ``the same curvature tensor'' as \((M_ 0,g_ 0)\). The authors prove that if \((M_ 0,g_ 0)\) is Lorentzian and irreducible, then \((M_ 0,g_ 0)\) has constant sectional curvature and each Lorentzian \((M,g)\) modelled on \((M_ 0,g_ 0)\) hsa also constant sectional curvature. On the other hand, they give examples of 3-dimensional Lorentz manifolds \((M,g)\) modelled on an indecomposable Lorentz symmetric space \((M_ 0,g_ 0)\) which are geodesically complete and not locally homogeneous.