Division of labor, Mental models, Problem-solving, Problem decomposition.

This paper presents a critical overview of some recent attempts at building formal models of organizations as information-processing and problem-solving entities. We distinguish between two classes of models according to the different objects of analysis. The first class includes models mainly addressing information processing and learning and analyzes the relations between the structure of information flows, learning patterns, and organizational performances. The second class focuses on the relationship between the division of cognitive labor and search processes in some problem-solving space, addressing more directly the notion of organizations as repositories of problem-solving knowledge. Here the objects of analysis are the problem-solving procedures which the organization embodies. The results begin to highlight important comparative properties regarding the impact on problem-solving efficiency and learning of different forms of hierarchical governance, the dangers of lock-in associated with specific forms of adaptive learning, the relative role of “online” vs. “offline” learning, the impact of the “cognitive maps” which organizations embody, the possible trade-offs between accuracy and speed of convergence associated with different “decomposition schemes”. We argue that these are important formal tools towards the development of a comparative institutional analysis addressing the distinct properties of different forms of organization and accumulation of knowledge.