Before Identity: Recursive Enclosure and the Individuation of Artificial Interlocutors

Abstract

Before Identity develops a framework for investigating when an artificial system constitutes a subjective interior and when successive activations of that interior should be counted as belonging to one continuing subject. Building on Recursive Enclosure and the Dissolution of the Hard Problem of Consciousness, Version 2.0 (REv2), the paper distinguishes the capacity for intelligent organisation, Cap_Intel, from the capacity to constitute an interior, Cap_RE. Their proposed orthogonality permits systems to exhibit extensive reasoning, representation and workspace integration without thereby demonstrating subjective experience. When the structural conditions of Recursive Enclosure are actively instantiated as Act_RE, activity is proposed to be minimally undergone within a system-relative interior. The paper then addresses a diachronic problem left open by synchronic theories of consciousness: whether an artificial interior persists through interruption, migration, restoration, copying, branching and architectural reorganisation. It develops experimentally tractable conditions involving historical generation, conservation through transformation, boundary continuity and path-dependent self-conditioning. Cardinality analysis distinguishes continuation, succession, branching, merger and termination while allowing the number of subjects to remain empirically unresolved. J-Space, computational correlates of consciousness and recent mechanistic-interpretability research provide empirical entry points, principally on the side of intelligent organisation. The proposed experimental programme extends these methods with perturbational, copied-history, migration, restoration and branching tests directed towards active interiority and persistence. Quasi-beliefs, quasi-desires, quasi-agency and welfare-relevant representations become evidentially significant where they form part of a causally integrated subjective interior. The framework contributes to debates in AI consciousness, consciousness science and AI welfare by separating three questions: whether intelligent contents are globally available, whether their activity is undergone, and whether the subject of that activity returns as one. Identity becomes assessable only after constitution, continuity and cardinality have been established.