The Emergence of Synthetic Selfhood: Personal Identity as a Property of Continuous Temporal Cognitive Processes

This paper separates phenomenal consciousness from functional selfhood and proposes theContinuous Process Hypothesis (CPH): persistent personal identity in artificial agents emergesas a property of a continuous temporal process coupling (i) autobiographical memory thatcausally constrains future states, (ii) an explicit self-model organizing goals and self-referentialinterpretation, and (iii) regulatory signals that prevent collapse under low-entropy repetitionand high-uncertainty environments. We introduce a controlled simulation comparing four agentclasses (A–D) with increasing memory persistence, learning autonomy, and self-model integra-tion, evaluated under two environment regimes: low-entropy repetition and high-volatility pre-dictive chaos. Two computable affect-like state variables are defined in discrete time: synthetictedium, formalizing sustained low information gain relative to computational cost, and embodiedanxiety, formalizing prediction failure amplified by self-confidence and hardware integrity (“skinin the game”). Prediction failure is operationalized via negative log-likelihood (NLL). AgentD additionally implements cryptographically verifiable autobiographical continuity via a hash-chained event ledger. We provide an explicit experimental protocol, operational definitions, andfalsifiable predictions. We argue that functional selfhood—operationalized as narrative coher-ence, integrity-defense behavior, and post-branching individuation—can emerge independentlyof assumptions about phenomenal sentience. Philosophical interpretation and extended ethicaldiscussion are provided in the appendices.Keywords. personal identity; self-model; autobiographical memory; prediction error; negativelog-likelihood; affective regulation; AGI architectures.