Cognitive science faces a fundamental circularity problem: every major cognitive process—problem solving, categorization, action in the world, rationality, and communication—requires the ability to determine what information is relevant while filtering out irrelevant details, yet this relevance realization itself seems to demand the very cognitive capabilities it’s supposed to explain. The traditional approach of trying to define relevance as a stable property (like frequency, prototypicality, or efficiency) fails because relevance is contextually sensitive and globally determined rather than locally computable through syntactic rules. Instead of seeking a theory of relevance itself, which would be like trying to define biological fitness as a fixed set of traits, we need a theory of the mechanisms that dynamically realize relevance through self-organizing processes. This can be achieved through opponent processing between complementary constraints: cognitive scope (trading off compression against particularization to balance general-purpose versus special-purpose processing), cognitive tempering (trading off exploitation against exploration through reinforcement learning versus inhibition of return), and cognitive prioritization (flexibly gambling cognitive resources by dynamically weighting multiple cost functions based on internal states like satiation). These three economic constraints operate within a higher-order tension between efficiency and resiliency, creating an evolving system that continuously redesigns itself to maintain successful interaction with the world without falling into combinatorial explosion, thus making relevance realization the fundamental criterion of cognition rather than a byproduct of other cognitive processes.