The importance of diversity in reasoning and learning to successfully address complex problems is examined. We discuss an approach by which a multiagent framework with decentralized control mechanisms provides diverse perspectives and hypotheses addressing a class of complex problems. We introduce the SNARE multiagent system. SNARE performs tasks to gain situational awareness of situations of interest in a Social Media Space. It applies a decentralized control mechanism for each agent; this mechanism enables an agent to interact with other agents to reason and learn. This approach facilitates dynamic agent organizations that adapt the topologies of interactions between agents based on the problem context.