Automatic concept evolution

Summary form only given, as follows. An evolutionary algorithm can be used by a distributed group of interacting people to produce the solution to a problem. First the problem must be defined. Each member of the group suggests solutions to the problem. They then go on to improve these suggestions by mutation and crossover. They show their preferences by replication and can throw in new suggestions if they think of them. A small experiment involving seven people using this algorithm to choose a name for it is described. The successes and limitations of this small experiment are discussed.