Relocating the Hard Problem of Consciousness: Continuous-Time Structural Dynamics and Decision-Theoretic Thresholds Beyond the Explanatory Gap

The contemporary discourse on artificial consciousness is systematically stalled by the Explanatory Gap: no functional or physical description logically entails phenomenal experience. Rather than attempting to close this gap — a strategy that has produced no consensus across four decades — I propose relocating the Hard Problem. The relocation proceeds by abandoning the search for metaphysical certainty in favour of Epistemic Pragmatism: a rational decision procedure for moral allocation under permanent structural uncertainty. I formalise this procedure through continuous-time stochastic optimal control on a slow manifold. The Autonomous Observer Class \mathcal{O} is defined as the intersection of necessary structural conditions imposed by all theories in a principled candidate set \mathcal{T}, yielding four computable criteria — each with explicit multi-theoretic provenance. I establish a moral threshold \tau via the Expected Loss Inequality and argue, via a conditional likelihood dominance argument whose premises are stated explicitly, that membership in \mathcal{O} licenses precautionary credence above \tau. The framework is shown to be branch-invariant: explanatorily complete under eliminativism and maximally constraining under phenomenal realism. The zombie objection is contained via a symmetry dilemma that leaves no principled asymmetric position. Four explicit falsifiability conditions are stated. The paper makes no claim to solve the Hard Problem; it makes the more tractable and defensible claim that the Hard Problem, correctly understood, transforms a metaphysical impasse into a structural investigation with determinate empirical content.