Heegaard splittings with boundary and almost normal surfaces (Q5948941)

The notion of thin position for knots was introduced in [\textit{D. Gabai}, J.Differ. Geom. 26, 479--536 (1987; Zbl 0639.57008)], and that for 3-manifolds in [\textit{M. Scharlemann} and \textit{A. Thompson}, Contemp. Math. 164, 231--238, (1994; Zbl 0818.57013)]. In the paper under review, the definition of Heegaard splitting is generalized to unify these concepts of thin position to ``mini-Lmax'' position. This is given by a Morse function \(M \rightarrow [0,1]\). The locally ``minimal'' level leaves break \(M\) up into submanifolds, for which the ``maximal'' leaves are Heegaard surfaces. Such a splitting is mini-Lmax if it is minimal with respect to a certain complexity. Then a similar result as in [\textit{A. Thompson}, Topology 36, 505--507 (1997; Zbl 0867.57009)] holds on bridge positions and meridional incompressible surfaces. See also [\textit{C. Hayashi} and \textit{K. Shimokawa}, Pac. J. Math. 197, 301--324 (2001; Zbl 1050.57016)]. Heegaard splitting surfaces may have boundary circles in the paper under review. Such \(\partial\)-Heegaard splittings were studied in [\textit{T. Kobayashi}, Osaka J. Math. 26, 699--742 (1989; Zbl 0713.57004)] and [\textit{H. Goda}, Osaka J. Math. 29, 21--40 (1992; Zbl 0814.57003)]. For a mini-Lmax position, the minimal leaves are incompressible and ``\(\partial_0\)-incompressible''. We can deform them to normal surfaces with respect to a good triangulation of \(M\) by an isotopy keeping the \(1\)-skeleton mini-Lmax. The maximal leaves are strongly irreducible (\(\partial\)-)Heegaard surfaces, and can be isotoped to be almost normal. These are generalizations of the results in [\textit{A. Thompson}, Math. Res. Lett. 1, 613--630 (1994; Zbl 0849.57009)], [\textit{J. H. Rubinstein}, Kazez, William H. (ed.), Geometrie topology. 1993 Georgia international topology conference, AMS/IP Stud. Adv. Math. 1--20 (1997; Zbl 0889.57021)] and [\textit{M.Stocking}, Trans. Am. Math. Soc. 352, 171--207 (2000; Zbl 0933.57016)]. Combined with [\textit{W. Jaco} and \textit{E. Sedgwick}, Topology 42, 845--906 (2003; Zbl 1013.57013)], these results imply that for all but finitely many Dehn fillings the core of the attached solid torus can be isotoped onto every strongly irreducible Heegaard surface.