Word representation of cords on a punctured plane (Q1763577)

For a set of \(n\geqslant 2\) points \(P_n = \{p_1, \ldots, p_n\}\) in \(\mathbb R^2\) define an \((i,j)\)-curve to be a continuous map \(\ell : [0,1] \to (\mathbb R^2 \setminus P_n) \cup \{p_i, p_j\}\) satisfying \(\ell(0) = p_i\) and \(\ell(1) = p_j\). Moreover, one assumes that \(\ell(p_i) = t\) if and only if \(t=0\) and that \(\ell(p_j) = 1\) if and only if \(t=1\). If the \((i,j)\)-curve \(\ell\) is simple, it is called an \((i,j)\)-cord, or simply a cord. The main theorem of the paper states that for a free group \(F_n\) there exists an explicitly computable map \(R_{ij} : F_n \to F_n\) such that (i) \(R_{ij}\) is a projection, namely \(R_{ij} \circ R_{ij} = R_{ij}\) and (ii) a word \(W\) in \(F_n\) is representable by an \((i,j)\)-cord if and only if \(R_{ij} (W) = (W)\). The computation of \(R_{ij}\) is purely algebraic. Although the study in the paper was confined to the case of a plane, similar arguments could be carried out to the 2-sphere. The motivation for this research is the monodromy problem appearing in Lefschetz fibrations and surface braids.