Combinatorial Lie bialgebras of curves on surfaces. (Q1430390)

The paper is devoted to a beautiful construction by Goldman and Turaev, which associates a bialgebra structure to the set of free homotopy classes of curves in two surfaces modulo trivial curves. The multiplication (comultiplication) of this algebra is defined by taking the sum over all smoothings (with respect to orientation) at intersection points with appropriate signs. For one curve, we take its self-intersection points and take the sum of skew-symmetric products of pairs of curves obtained by such smoothings, for two curves we take the sum of all curves obtained from this curve by smoothing at all vertices where one curve intersects the other one. Clearly, the product of two non-intersecting curves gives zero, as well as the co-product of a simple curve. Thus, there are several natural conjectures concerning the dependence of the minimal number of intersection points for curves in a given class (resp., self-intersection points for one curve) and the number of terms in the Lie bracket of this pair of curves (resp., Lie cobracket for the curve). The author gives a nice algebraic interpretation of the Goldman-Turaev bialgebra in terms of cyclic words related to the generators of the fundamental group of the surface. This allows to solve some previous conjectures by Turaev, establish new properties and state some new conjectures concerning intersection points and terms in the bracket/cobracket expansion.