Abstract

Abstract This paper formalizes a claim that runs implicitly through my earlier work in recursive epistemic engineering: information is not fundamentally stored in a substrate; information is stabilized recursion. The substrate--biology, silicon, social infrastructure, or field-level physics--matters only insofar as it supplies boundaries that allow recursion to persist under drift. I define the recursive substrate of information as a minimal operator stack that repeatedly reappears whenever a system must: 1. continue itself, 2. remain coherent across scale, and 3. resist runaway instability. This yields a constraint vocabulary: (\Phi,\; \mathfrak{F},\; \nabla H) where Phi is continuation through re-entry, F is scale-continuity (fractal coherence), and nabla H is stabilization pressure (entropy / drift gradient). I treat this not as metaphysical finality but as an engineering-and-critique scaffold: a bridge-language validated by load rather than belief. The model is falsifiable. If recursive attractors do not converge under pressure, or if cross-observer semantic structures do not exhibit stability under controlled protocols, the framework must be revised. The McPhetridge Experiment is introduced as a bounded empirical anchor: a convergence test for whether informational structure collapses differently per observer or persists as an externally stabilized attractor within the public semantic layer. Implications extend across AI boundary dynamics, cognition, and informational cosmology, but the claim remains scoped: the substrate of information is not matter--it is the loop that carries itself. --- 1. Introduction -- Why Substrate Debates Miss the Primitive The debate over "substrate dependence" in mind, AI, and information theory tends to take a familiar form: * One camp argues functional equivalence: if the computation is the same, the substrate doesn't matter. * Another camp argues embodiment constraints: thermodynamics, biology, or physics are essential. Both camps are usually arguing at the wrong layer. The deeper question is not which substrate carries information, but: What makes information persist at all? What makes it remain coherent under drift? What makes a pattern survive perturbation? In my framework work (Phi-Method, REM, and the triad papers), the core move is always the same: * identity is not a stored object * truth is not a static statement * information is not a passive bitstring All three are stable attractors of recursive processes constrained by failure laws. This paper extends that move explicitly: it treats recursion itself as the "substrate" of information. The physical medium is downstream. The primary carrier is a stability-preserving loop.