Modeling Structure and Cognition in Organizations: A Meta-Network Computational Approach

Introduction In this paper we present a meta-network approach to the analysis of organizations in terms of three basic domain elements: cognitive agents, tasks and resources (Krackhardt and Carley, 1998). Building on mainstream social network and distributed artificial intelligence literature, we propose that formalizing dependencies between domain elements at various levels provides a rich grammar for theorizing about organizations. We demonstrate the versatility and utility of this approach for generating a series of testable hypotheses about organizational processes and performance. These hypotheses are tested through a series of agent-based social simulations. It is shown that both structure and agents’ cognition are important predictors of how an organization works, and what results it achieves. In turn, the predictive power of structure and cognition depends on the generation of group mind-like forms of mental models emerging from a network of socially and cognitively integrated agents.