Some splitting theorems for stably causal space-times (Q1267042): Difference between revisions

Let \((M,g)\) be a connected spacetime which satisfies the strong energy condition, \(\text{Ric}(X,X)\geq 0\) for timelike vectors \(X\), the celebrated Lorentzian splitting theorem states that if \((M,g)\) is timelike geodesically complete and contains a timelike line, then \((M,g)\) is isometric to a product of the timelike line and a complete Riemannian manifold. The key idea in the proof is to use the Busemann functions associated with the timelike line. The paper under review replaces the existence of a timelike line by the existence of a special function which plays essentially the same role as the Busemann function. To be more precise, the authors prove that if a connected space time \((M,g)\) admits a smooth function \(f\) with the property that \(g(\nabla f,\nabla f)<0\), \(\Delta g(\nabla f,\nabla f) \geq 0\), \(\text{Hess}(f)\) is diagonalizable at each point, and \(\text{Ric}(\nabla f,\nabla f)\geq 0\), then \((M,g)\) is isometric to a product of a timelike line and a Riemannian manifold, provided that \(\nabla f\) is a complete vector field. Another related theorem proved in the paper is the following: If \((M,g)\) admits a function \(f\) with the property that \(g(\nabla f,\nabla f)=-1\), and \(\text{Ric}(\nabla f,\nabla f)\geq 0\), then \((M,g)\) is isometric to a product of a timelike line and a complete Riemannian manifold, provided that \((M,g)\) is finitely compact. In the note added in proof, the authors discuss the condition of finite compactness.