Shadows, Simulations and Hallucinations
This thesis argues that hallucination is not a cognitive error but an essential generative function of sophisticated cognitive systems, both biological and artificial. It explores historical and philosophical foundations, including Plato's Allegory of the Cave, Descartes' radical skepticism, and Schrödinger's observer-dependent reality. A central empirical focus is the 1970s Stanford Research Institute remote viewing program, funded by the CIA, which used double-blind methodologies and produced mixed results—sometimes accurate, sometimes convincingly inaccurate. The thesis contends that producing credible yet unverifiable representations is an adaptive strategy for navigating uncertain environments, and recognizing hallucination as fundamental to intelligence suggests new evaluation criteria for AI safety and insights into consciousness research.