From Hive Minds to Administered Minds: A Formal Theory of Agency Collapse and Emergent Macro-Agency

In Lerer (2025a), I argued that game theory's foundational assumptions formalize the structural conditions of human agency. Here I formalize this intuition by establishing an equivalence with hive mind dynamics. Drawing on Soma et al.'s (2025) proof that populations following imitation rules are mathematically equivalent to single reinforcement learning agents, I show that the four conditions of strategic agency (separation, imperfect information, bounded rationality, minimal symmetry) map onto the structural prerequisites for maintaining individual agency within collective systems. When these conditions erode—through technological transparency, algorithmic coordination, or institutional design—individual agents collapse into a macro-agent: a "hive mind" or "administered mind." The formal equivalence means this is not metaphor but mathematical identity. Corporate structures, social media platforms, and certain governance regimes produce administered minds by design, transforming moral agents into optimization outputs of a higher-level entity. The paper tries to provide a formal framework for understanding agency collapse and identifies institutional safeguards that preserve individual autonomy. This has implications for constitutional design, antitrust policy, AI governance, and theories of personhood.