A local calculus for nullhomotopic filling Dehn spheres (Q1017855)

Let \(M\) be a closed 3-manifold. A subset \(\Sigma\) of \(M\) is a Dehn sphere if there is a transverse immersion \(f:S^2 \rightarrow M\) such that \(\Sigma=f(S^2)\). \(\Sigma\) is called filling if its singularities induce a cell decomposition of \(M\), that is, if \(S(\Sigma)\) denotes the singular set of \(\Sigma\) and \(T(\Sigma)\) denotes the set of triple points, then \(T(\Sigma)\not= \emptyset\), \(S(\Sigma)\setminus T(\Sigma)\) consists of intervals, \(\Sigma\setminus S(\Sigma)\) consists of disks and \(M\setminus\Sigma\) consists of 3-balls. \(\Sigma\) is nullhomotopic if the map \(f\) defining it is homotopic to a constant map. It is known by the work of \textit{J. M. Montesinos-Amilibia} [Mathematical contributions. Homage to Prof. D. Joaquín Arregui Fernández. Madrid: Editorial Complutense. Homenajes de la Universidad Complutense. 239--247 (2000; Zbl 0981.57009)] that any closed 3-manifold admits a nullhomotopic filling Dehn sphere. In the paper under review a local calculus for nullhomotopic filling Dehn spheres is given. That is, it is proved that if \(\Sigma_1\) and \(\Sigma_2\) are nullhomotopic filling Dehn spheres of closed 3-manifolds \(M_1\) and \(M_2\) respectively, then \(M_1\) and \(M_2\) are homeomorphic if and only if there is a sequence of certain local moves that transform \(\Sigma_1\) in \(\Sigma_2\). Another set of moves has been given previously by \textit{R. Vigara} [Fundam. Math. 190, 245--288 (2006; Zbl 1100.57023)], but his moves are somehow non-local. To prove the main result, the author shows that any of Vigara's moves can be realized by a sequence of the new moves. Given a nullhomotopic filling Dehn sphere, the author defines a state sum invariant, which is a polynomial of several variables modulo a certain ideal, which turns out to be invariant under the given moves and then defines an invariant of closed 3-manifolds. Such an invariant is defined in an analogous manner as the invariant of \textit{V. Turaev} and \textit{O. Viro} [Topology 31, No.~4, 865--902 (1992; Zbl 0779.57009)], but as the author says, it is unclear whether or not both invariants are related.