Creative Problem Solving Groups at the Edge of Chaos

Mathematical concepts of chaos and self‐organization were used to reinterpret the chance‐configuration process for creative problem solving. Transcripts from three problem solving discussions involving eight or fewer social scientists were analyzed by nonlinear regression to determine whether the groups' productivity was chaotic over time and if control variables were in evidence. It was shown that productivity was chaotic as evidenced by a positive Lyapunov exponent, strongly correlated with the number of active discussion threads, and also dependent on the people‐discussion combination. The number of discussion threads, in turn, acted as a driver function that triggered chaosticity in overall performance. The results suggested numerous new studies for investigating dynamical processes in creative thought.