The Physics of Belief: Defining the Wimberly Limit in Biological and Synthetic Intelligence

For centuries, science has treated information as massless and logic as frictionless. We assume that when facts change, minds change instantly. This paper presents a fundamental correction: Information has Inertia. Introducing The Theory of Cognitive Inertia, this research unifies findings from Large Language Models (AI), exercise physiology, and clinical pathology to prove that "Truth" is a thermodynamic state. We demonstrate that the act of changing a category—whether in a neural network or a human brain—requires a specific amount of metabolic work. We identify The Wimberly Limit (Lambda-W): a critical metabolic threshold. When a system's energy drops below this limit, it becomes physically impossible to update a belief, resulting in "Reasoning Rigidity" in AI and "Dogmatism" in humans. By quantifying this Viscosity Index, we resolve the Sorites Paradox not as a linguistic failure, but as an energy-conservation strategy. We conclude that "Stubbornness" is not a character flaw; it is a calculated defense against metabolic ruin.