A Structural-Dynamical Account of Consciousness and Self-Awareness

Existing frameworks for analyzing consciousness rely on single-scalar measures, substrate-specific assumptions, or insufficient evidential discipline for cross-system comparison. This working draft proposes a substrate-neutral structural-dynamical framework that addresses all three limitations. Consciousness is defined as integrated, recursive, self-referential world modeling with temporally extended trajectory weighting. The framework separates architectural capacity from realized conscious episodes and introduces an eight-coordinate profile vector C(S, t) = ⟨I, D, T, Sₑ, B, Eᴄ, K, V⟩ for comparative, non-scalar characterization of systems across biological and artificial substrates. Self-awareness is treated as graded region membership within this multidimensional space. The structural characterization is paired with a perturbation-family evidence governance layer—structured rules specifying what counts as admissible evidence for placement updates, including evidence tiers, axis-local revision rules, and safeguards against premature attribution. Phenomenology claims are treated as defeasible bridge principles subject to structured revision. The framework yields a concrete research program in which deployment-architecture variation—persistent memory, tool use, planning loops—serves as the primary source of structural contrast, governed by the perturbation-family policy.