Conflict escalation will be approached from a perspective of Dynamical Systems Theory. It is proposed that escalation and de-escalation of conflict result in changes of the ensemble of possible behaviors. We employ a combination of multidimensional scaling and network analysis to understand the dynamics of change of possible states accompanying changes in conflict intensity. The network is characterized by attractor driven dynamics in the sense that it converges in one of several possible configurations. This model represents one of possible interpretations of conflict as an attractor.