The Dual Structure of Mind and Life Can Explain Consciousness and Evolutionary Creativity Beyond Simulation

We propose that complex systems in mind and life follow a dual architecture: a computational component generating structures (e.g., thoughts, biochemical states) and a directional component ensuring long-term coherence and creativity. Large language models (LLMs) simulate the former but lack the latter, resulting in entropic decay during autonomous operation—a process governed by Shannon’s Data Processing Inequality. Drawing on autopoiesis theory*and developmental canalization, this hypothesis posits that consciousness and evolutionary directionality are ontologically distinct control layers, falsifiable through long-term AI autonomy tests and evolutionary simulations. The framework bridges empirical AI failures with theoretical principles of adaptive complexity, offering testable predictions for complex systems research.