Hyperbolic manifolds, harmonic forms, and Seiberg--Witten invariants (Q1611627)

The present paper circles around the (very well-motivated) conjecture that any compact \(4\)-dimensonial anti-self-dual Einstein manifold with negative scalar curvature has vanishing Seiberg-Witten invariants (for the fixed orientation). A recent result of \textit{J. Armstrong} [Q. J. Math., Oxf. (2)48, 405-415 (1997; Zbl 0899.53038)] asserts that anti-self-dual Einstein spaces of negative scalar curvature never admit nontrivial self-dual harmonic \(2\)-forms of constant length. On the other hand, the present article shows that if such a space has a non-trivial Seiberg-Witten invariant, it necessarily admits a self-dual harmonic \(2\)-form whose length is ''nearly constant''. To this end, the author introduces two different measures of ''constantness'' and defines corresponding invariants \(\theta, \nu\) which offer quantitative obstructions to the existence of nontrivial self-dual harmonic \(2\)-forms of constant length on a given \(4\)-manifold. He then deduces several new Seiberg-Witten estimates relating these invariants to standard topological invariants. One corollary is that a compact \(4\)-dimensonial anti-self-dual Einstein manifold with scalar curvature \(s=-12\), \(b^+\geq 2\) and \(\nu\geq 2-\sqrt{3}\) has vaninshing Seiberg-Witten invariants for the given orientation. The question how large this class of manifolds actually is, has to remain open.