An arithmetic Zariski 4-tuple of twelve lines (Q258900)

A celebrated result due to Zariski from the thirties shows that the topology of the embedding of an algebraic curve \(C\) into the complex projective plane is not determined by the local topology of its singularities. Namely, he used two different sextics with six cusps in order to show that the fundamental groups of their complements are not isomorphic. Such a pair of curves is called a Zariski pair and this notion was introduced in the middle of the nineties by \textit{E. Artal Bartolo} [J. Algebr. Geom. 3, No. 2, 223--247 (1994; Zbl 0823.14013)]. Now we know that there are some examples of Zariski pairs, but we have only two examples of such pairs given by line configurations in the complex projective plane, namely, the MacLane configurations [\textit{G. L. Rybnikov}, Funct. Anal. Appl. 45, No. 2, 137--148 (2011; Zbl 1271.14085); translation from Funkts. Anal. Prilozh. 45, No. 2, 71--85 (2011)] and the configurations constructed quite recently by \textit{E. Artal} et al. [Compos. Math. 141, No. 6, 1578--1588 (2005; Zbl 1085.32012)]. In the present note the author finds a new (arithmetic) Zariski 4-tuple determined by line configurations of twelve lines. Let us recall some fundamental definitions in order to present the main results of the paper. An abstract combinatorics is a pair \(\mathcal{C} = \{ \mathcal{L}, \mathcal{P} \}\), where \(\mathcal{L}\) is a finite set and \(\mathcal{P}\) is a subset of the power set of \(\mathcal{L}\), satisfying \(\#P \geq 2\) for all \(P \in \mathcal{P}\) and, for any \(L_{1}, L_{2} \in \mathcal{L}\) with \(L_{1} \neq L_{2}\), there exists a unique \(P \in \mathcal{P}\) such that \(L_{1}, L_{2} \in P\). We say that the combinatorics \(\mathcal{C}\) is ordered if the set \(\mathcal{L}\) is ordered. For any abstract combinatorics \(\mathcal{C}\) we say that \(\tau\) is an automorphism of \(\mathcal{C}\) if it is a permutation of \(\mathcal{L}\) preserving \(\mathcal{P}\). Let \(\mathcal{L}\) be a line configuration, then the combinatorics of \(\mathcal{L}\) is the poset of all intersections of the elements in \(\mathcal{L}\) with respect to the revers inclusion. For a given (abstract) combinatorics \(\mathcal{C}\), a complex line configuration \(\mathcal{L}\) is a realization of \(\mathcal{C}\) if its combinatorics agrees with \(\mathcal{C}\). An ordered realization of an ordered combinatorics is defined in the same spirit. Finally, the topological type of a line configuration \(\mathcal{L}\) is the homeomorphism type of the pair \((\mathbb{P}^{2}_{\mathbb{C}}, \mathcal{L})\). If the homeomorphism preserves the orientation of \(\mathbb{P}^{2}_{\mathbb{C}}\), then we have an oriented topological type, and it is ordered if \(\mathcal{L}\) is ordered and the homeomorphisms preserve this order. It is worth to point out here that two MacLane configurations have the same combinatorics, but they have different oriented and ordered topological type, thus they are in distinct ambient isotopy classes. The main results deals with four line configurations of eleven lines (which are not free in the sense of Saito) having exactly \(4\) quadruple points, \(6\) triple points and \(13\) double points defined by slightly different equations -- the difference is made by a choice of two roots of the \(10\)-th cyclotomic polynomial and just by taking (complex) conjugations of these two roots. Denote these configurations by \(\mathcal{L}_{1}, \widetilde{\mathcal{L}}_{1}, \mathcal{L}_{2}, \widetilde{\mathcal{L}}_{2}\). Theorem 1. There is no homeomorphism preserving both orientation and order between any two pairs among \((\mathbb{P}^{2}_{\mathbb{C}}, \mathcal{L}_{1})\), \((\mathbb{P}^{2}_{\mathbb{C}}, \mathcal{L}_{2})\), \((\mathbb{P}^{2}_{\mathbb{C}}, \widetilde{\mathcal{L}}_{1})\), \((\mathbb{P}^{2}_{\mathbb{C}}, \widetilde{\mathcal{L}}_{2})\). Theorem 2. There is no order-preserving homeomorphism between \((\mathbb{P}^{2}_{\mathbb{C}}, \mathcal{L}_{1})\) and \((\mathbb{P}^{2}_{\mathbb{C}}, \mathcal{L}_{2})\), or between \((\mathbb{P}^{2}_{\mathbb{C}}, \mathcal{L}_{1})\) and \((\mathbb{P}^{2}_{\mathbb{C}}, \widetilde{\mathcal{L}}_{2})\), or between \((\mathbb{P}^{2}_{\mathbb{C}}, \widetilde{\mathcal{L}}_{1})\) and \((\mathbb{P}^{2}_{\mathbb{C}}, \mathcal{L}_{2})\), or between \((\mathbb{P}^{2}_{\mathbb{C}}, \widetilde{\mathcal{L}}_{1})\) and \((\mathbb{P}^{2}_{\mathbb{C}}, \widetilde{\mathcal{L}}_{2})\). In order to show these results the author studies a new topological invariant \(\mathcal{I}(\mathcal{L}, \xi, \gamma)\) of line configurations introduced together with \textit{B. Artal} and \textit{V. Florens} in [``A topological invariant of line arrangements'', Preprint, \url{arXiv:1407.3387}], which depends only on the character of the fundamental group of the complement and a special cycle in the Levi graph of the configuration. It turns out that this invariant is different for each configuration, which allows to conclude the results. It is worth to notice that the proposed invariant can be read off using the so-called braided wiring diagrams (please consult Section 4). In order to get a new Zariski pair one needs to get rid of automorphisms of considered line configurations and this can be done by adding in a clever way one line in such a way that the new combinatorics consists of one quintuple point, \(3\) quadruple points, \(6\) triple points and \(20\) double points. The paper contains new really interesting ideas and is pleasant to read. For more details the reviewer encourages also to look at the PhD thesis of the author.