Common sense suggests that work is organized in accordance with plans that are created by designers who reflect on the work setting and manipulate representations of the work process in order to determine new and efficient organizational structures. Or, even if "outside" designers are not involved, the reorganization of work is normally attributed to the conscious reflection by members of the work group itself. A detailed examination of the response of a real-world group to a sudden and unexpected change in its informational environment shows that these common sense assumptions may be quite misleading. While entering a harbor, a large ship suffered an engineering breakdown that disabled an important piece of navigational equipment. This paper considers the response of the ship's navigation team to the changed task demands imposed by the loss of this equipment. Following a rather chaotic search of the space of computational and social organizational alternatives, the team arrived at a new stable work configuration. In retrospect, this solution appears to be just the sort of solution we would hope designers could produce. However, while some aspects of the response appear to be the products of conscious reflection, others, particularly those concerning the division of cognitive labor, are shown to arise without reflection from adaptations by individuals to what appear to them as local task demands. It is argued that while the participants may have represented and thus learned the solution after it came into being, the solution was clearly discovered by the organization itself before it was discovered by any of the participants.
