On six-dimensional pseudo-Riemannian almost g.o. spaces (Q2456446)

A (connected) pseudo-Riemannian manifold \((M,g)\) is homogeneous if there exists a group \(G\) of isometries acting transitively on it. Such \((M,g)\) can be identified with \((G/H,g)\), where \(H\) is the isotropy group at a fixed point \(o\) (the origin) of \(M\). The pair \((G/H,g)\) is called a pseudo-Riemannian homogeneous space. If the metric \(g\) is positive definite, then \((G/H,g)\) is always a reductive homogeneous space; this means that, denoting by \(\mathfrak{g}\) and \(\mathfrak{h}\) the Lie algebras of \(G\) and \(H\), respectively, there exists a subspace \(\mathfrak{m}\) of \(\mathfrak{g}\) such that \(\mathfrak{g}=\mathfrak{h}\oplus \mathfrak{m}\) and \(\text{Ad}(H)(\mathfrak{m})\subset \mathfrak{m}\) (where \(\text{Ad}: H\times \mathfrak{g} \rightarrow \mathfrak{g}\) is the adjoint representation of \(H\) on \(\mathfrak{g}\)). If the metric \(g\) is indefinite, the reductive decomposition of the Lie algebra \(\mathfrak{g}\) may not exist [see \textit{J. Figueroa-O'Farril, P. Meessen} and \textit{S. Philip}, Homogeneity and plane wave limits, arXiv: hep-th/0504069 (2005)]. Let \((G/H,g)\) a pseudo-Riemannian homogeneous space, \(\mathfrak{g}=\mathfrak{h}\oplus \mathfrak{m}\) a reductive decomposition. A geodesic \(\gamma(s)\) through the origin \(o\) is homogeneous if there exists a reparametrization \(\gamma^*(t)\) of \(\gamma(s)\) such that \(\gamma^*(t)=\exp(tX)(o)\) for any \(X\in \mathfrak{g}\). The vector \(X\) is called a geodesic vector. A pseudo-Riemannian homogeneous space \((G/H,g)\) is called a g.o. space (g.o. means geodesics are orbits) if all geodesics of \((G/H,g)\) are homogeneous. It is well known that all naturally homogeneous spaces are g.o. spaces but the converse is not true. The first example of a g.o. space which is in no way naturally reductive is due to A. Kaplan. This is a six-dimensional Riemannian nilmanifold with two-dimensional center, one of the so-called generalized Heisenberg groups. In a previous paper [Examples of pseudo-Riemannian g.o. manifolds, Proc. 8th Int. Conf. Varna 2006, 145--155 (2007; Zbl 1122.53030)], the authors gave examples of five-dimensional, six-dimensional and seven-dimensional pseudo-Riemannian g.o. spaces with compact isotropy group. They also described the discontinuities of geodesic graphs. At all points of discontinuity, there are different limits along different curves, but these limits are all finite. In the paper under review, the authors modify the Riemannian metrics on the six-dimensional Kaplan's g.o. manifold and obtain homogeneous pseudo-Riemannian manifolds with non compact isotropy group. They describe the (incomplete) geodesic graphs on open dense subset \(U\) of \(\mathfrak{m}\), but they show that these homogeneous pseudo-Riemannian manifolds are not g.o. manifolds. They also prove that the (incomplete) geodesic graphs are strongly discontinuous at the boundary, i.e., the limits along certain curves are always infinite.