As humans, we tackle many problems in complex societies and manage the complexities of networked social systems. Cognition and sociability are two vital human capabilities that improve social life and complex social interactions. Adding these features to smart devices makes them capable of managing complex and networked Internet of Things (IoT) settings. Cognitive and social devices can improve their relationships and connections with other devices and people to better serve human needs. Nowadays, researchers are investigating two future generations of IoT: social IoT (SIoT) and cognitive IoT (CIoT). This study develops a new framework for IoT, called CSIoT, by using complexity science concepts and by integrating social and cognitive IoT concepts. This framework uses a new mechanism to leverage the friendships between devices to address service management, privacy, and security. The framework addresses network navigability, resilience, and heterogeneity between devices in IoT settings. This study uses a new simulation tool for evaluating the new CSIoT framework and evaluates the privacy-preserving ability of CSIoT using the new simulation tool. To address different CSIoT security and privacy issues, this study also proposes a blockchain-based CSIoT. The evaluation results show that CSIoT can effectively preserve the privacy and the blockchain-based CSIoT performs effectively in addressing different privacy and security issues.