This paper defines Co-Cognition as a core concept within the broader Reality Drift framework, describing the distributed cognitive process through which human thought and external systems, particularly AI systems, operate in shared feedback loops to generate, refine, and apply representations. Rather than functioning as passive tools, external systems increasingly participate in the recursive production of cognition itself. In this process, outputs from human and machine systems become inputs for subsequent stages of thought, creating distributed reasoning loops that extend cognition across internal and external layers. The paper examines the structural mechanisms underlying co-cognitive systems, including recursive compression, semantic fidelity, cognitive grounding, and constraint preservation. As a canonical concept paper within the Reality Drift framework, it provides the primary definitional basis for understanding how cognition becomes distributed across human and artificial systems and how drift can emerge when those shared loops lose alignment with reality.