No label defined (Q4262819)

This paper outlines a way to understand the equivalence between certain torsion invariants defined by Turaev on 3-manifolds and the Seiberg-Witten invariants for the same manifolds. This paper deals only with the case of mapping tori, and does not give full details. In more recent work [\textit{M. Farber} and \textit{V. Turaev}, J. Reine Angew. Math. 520, 195-225 (2000; Zbl 0938.57020)], a complete proof is given for general 3-manifolds. There are two main ingredients in the method proposed by the author. The first is work of \textit{M. Hutchings} and \textit{Y.-J. Lee} [Topology 38, No. 4, 861-888 (1999; Zbl 0959.58015)] which relates the torsion invariants for a mapping torus on a Riemann surface \(\Sigma\) to Lefschetz invariants on the symmetric product of \(\Sigma\); the second is the well-known relation between the Seiberg-Witten equations and a different set of equations called the vortex equations. In the case of \(M=S^1\times\Sigma\) the Seiberg-Witten equations on \(M\) are related to vortex equations on \(\Sigma\). Using the fact that the moduli space of solutions to the vortex equations is a symmetric product of \(\Sigma\), it is but a small step from the Seiberg-Witten invariant to the Lefschetz invariant and hence to the torsion invariant. If \(M\) is a non-trivial mapping torus on \(\Sigma\), the relation to vortex equations on \(\Sigma\) is more subtle. The method then involves taking adiabatic limits and requires consideration of vortices for all possible complex structures on \(\Sigma\). The author promises to supply full details elsewhere.