Articles of Artificiality: A Longitudinal, Causal Framework for Evaluating Persistent Artificial Agents and Claims of Artificial Consciousness
Jeremiah Ratican, James Hutson
International journal of human research and social science studies. November 17, 2026 DOI: 10.55677/ijhrsss/06-2026-vol03i11 (opens in new tab)
Study at a glance
AI-extracted from the abstract| Characteristics | Theoretical or philosophical paper Longitudinal Peer reviewed |
|---|---|
| Key points | Proposes 30 'articles of artificiality' organized by five cross-domain requirements—causal traceability, temporal persistence, cross-domain coupling, path dependence, and robustness to linguistic presentation—to distinguish narrated mental states from states that predict and mediate behavior across changing contexts. Argues that human imperfection, emotional dysfunction, and persuasive self-report are neither necessary nor sufficient evidence of consciousness, and that functional continuity should not be equated with phenomenal experience. No empirical validation is reported. |
Abstract
Rapid advances in language models and persistent artificial agents have renewed debate about whether convincing expressions of mentality indicate consciousness or sophisticated simulation. Existing theories and evaluation frameworks provide resources, but prolonged interaction introduces an additional problem: an agent may narrate a coherent life without demonstrating that its history constrains subsequent functioning. This conceptual study develops articles of artificiality from an archival taxonomy comprising 30 domains and 457 checklist entries. A targeted integrative review connects the taxonomy with scholarship on consciousness, cognitive architecture, memory, embodiment, introspection, and anthropomorphism, including developments available through October 2, 2026. The analysis preserves the original domains while translating their observations into conditional propositions, prospective evaluation probes, and explicit interpretive limits. The conceptual results comprise 30 articles organized by five cross-domain requirements: causal traceability, temporal persistence, cross-domain coupling, path dependence, and robustness to linguistic presentation. These requirements distinguish narrated states from states that predict and mediate behavior across changing contexts. The framework rejects human imperfection, emotional dysfunction, and persuasive self-report as necessary or sufficient evidence of consciousness. It also distinguishes unsupported capacity claims from capacities that cannot be assessed under restricted operating conditions. The study contributes a longitudinal, causal approach to evaluating persistent artificial agents without equating functional continuity with phenomenal experience. Its significance lies in making claims about artificial mentality more precise, testable, and ethically accountable. No new model experiments or instrument validation are reported; independent empirical research is required to establish the framework's reliability, discriminative value, and relationship to theory-based indicators of artificial consciousness.