Agency and Self-Awareness in Phase State Geometry (PSG): A Formal Framework for 4th-Order Recursive Stability
Year: 2026
Type: article
Abstract: This document serves as the theoretical bridge between the foundational axioms of Phase State Geometry (PSG) Books I-IV and the practical implementation of stable, self-modeling Artificial General Intelligence. It formalizes the 4th-Order Recursive Architecture required to transition from simple data representation to autonomous self-simulation. Central to this framework is the Refinement Lagrangian, a variational principle that unifies agency and self-awareness by demonstrating that intelligence naturally acts to minimize 'structural tension' across internal and external models. By applying this principle, the paper provides a mathematical blueprint for achieving Refinement-Stable Attractors, defining the 'Self' as a fixed-point invariant within a joint-stable manifold and offering a rigorous, geometric path toward AGI alignment.
Author: Renaud, Albert
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7160761501
FWCI:
Cites:
Cited by:
Related to: