In order to apply control mechanisms found in biological phenomena to social design, K. Gonda developed a theory of "fields" which simulates the process of knowledge creation in a community as a dynamically stable system involving self-organization under multiple controls. This theory is an original type of cluster theory, which focuses attention to the spillover of knowledge in a community. Meanwhile, in other countries, the network analysis has been developing which focuses on the network of actors in a particular region as a knowledge generation system, while the recent evolutionary economics argues that the flow of knowledge in a cluster is selective and heterogeneous depending on the relational distance between enterprises in the cluster, thus criticizing econometric assumption that emphasizes the role of spatial distance in the flow of knowledge. Thus the cluster theory in Europe and America has overcome simplistic knowledge spillover model based on the spatial distance between the actors and proceeded to an approach based on the system theory. The current studies by the author on clusters from the viewpoint of network analysis have led him to reaffirm Gonda's foresight incorporated in his research agenda.