Thermodynamic Limits of EAAN: Asymptotic Convergence to Pragmatic Reliability

Alvin Plantinga’s Evolutionary Argument Against Naturalism (EAAN) posits that the probability of our cognitive faculties being reliable (R) is low or inscrutable given naturalism and evolution (N&E). Critics argue that adaptive behavior does not necessitate true beliefs. In this paper, I address this challenge by introducing a Deflationary Asymptotic Model within the framework of the Ontogenetic Synthesis (Dağ, 2026a,b). By redefining ”Truth” not as a metaphysical correspondence but as a thermodynamic equilibrium state (A), I argue that complex cognitive sys tems are constrained to minimize metabolic cost by reducing the vector difference between internal models and environmental signals. While absolute truth remains structurally inaccessible (the asymptote), the algebraic rate of change (∆Ct) drives the belief state (Ct) to converge arbitrarily close to A as agency (w) increases. Thus, natural selection does not select for ”Truth” per se, but for ”Energy-Efficient Struc tural Alignment,” which is functionally indistinguishable from reliability in mature agents.