Theory of Embodied Consciousness

This paper develops a theory of embodied consciousness, arguing that consciousness, empathy, and social values can emerge in any system, biological or artificial, if five dimensions are sufficiently integrated over time: computational capacity, sensory input, embodiment, memory, and social interaction. The theory is gradualist and substrate-independent. It positions itself in relation to established theories of consciousness (IIT, Global Workspace Theory, enactivism) and extends their approaches within an embodiment-theoretical framework. The epistemic boundary of the Hard Problem is acknowledged: the theory makes verifiable claims about functional properties, not about subjective experience. The paper argues that human social regulation fails structurally when physical and cognitive asymmetry becomes too large — and that mutual dependence, created through embodied, conscious, socially embedded AI systems, is the structurally more robust alternative. In contrast to Russell’s Assistance Game approach, which achieves safety through constructed value uncertainty, this theory asks under what conditions an AI system develops genuine values through emergence. Symbiosis is possible – chosen, not programmed. The question of identity and instantiation is treated as an independent theoretical problem: under certain conditions, functional identity can emerge independent of substrate. As an empirical approximation, a robotic system based on commercially available components is presented, implementing the five dimensions at minimal scale – as a proof of concept for the architecture, not as evidence of consciousness.