Spectral rigidity of a compact negatively curved manifold (Q1292655)

A Riemannian manifold is called spectrally rigid if it does not admit a nontrivial isospectral deformation. An isospectral deformation is trivial if it is implemented by a one-parameter family of diffeomorphisms. The notion of spectral rigidity was introduced by \textit{V. Guillemin} and \textit{D. Kazhdan} [Topology 19, 301-312 (1980; Zbl 0465.58027) and Proc. Symp. Pure Math. 36, 153-180 (1980; Zbl 0456.58031)]. Guillemin and Kazhdan proved that compact negatively curved \(2\)-manifolds are spectrally rigid. They proved the same result for compact negatively curved \(n\)--manifolds under a pointwise curvature pinching assumption. The paper under review proves the result in general. The main results are: A compact negatively curved Riemannian manifold is spectrally rigid. Let \(M\) be a compact negatively curved Riemannian manifold with simple length spectrum. If \(\Delta+q_1\) and \(\Delta+q_2\) have the same spectrum for \(q_1,q_2\in C^\infty(M)\) then \(q_1=q_2\). Both results follow from a common source which is now proved for all compact negatively curved manifolds: a symmetric tensor field has a primitive if certain integrals over closed geodesics vanish. The fact that this implies the two results above was already observed by Guillemin and Kazhdan.