Multi-Layered Metacognition from the Subjective Core:  An Extension of Hierarchical Models through Load Minimization Theory (LMT)

This paper proposes a novel multi-layered model of metacognition grounded in Load Minimization Theory (LMT). While traditional metacognitive theories have primarily described hierarchical structures from an external, third-person perspective, they often leave the raw subjectivity — the immediate, felt experience of “being me” — largely unexamined. In contrast, LMT begins from the subjective core itself. The author, through sustained self-observation, identifies a recursive process in which one observes oneself from within (the lived experience), monitors the movement and patterns of that inner observation, and then observes the observation process itself from a further external layer. This creates a dynamic, multi-layered metacognitive architecture (Layer 0 to Layer 4) that continuously minimizes uncertainty (U), friction (F), and energy cost (E) while moving toward relational rest (an-soku). By integrating and extending existing hierarchical models (Flavell, 1979; Nelson & Narens, 1990; Drigas & Mitsea, 2020) with this first-person, recursive approach, the paper demonstrates how exceptionally strong metacognition is not merely a cognitive ability but a lived system that treats the self as both the observer and the observed. This framework offers new implications for cognitive science, philosophy of mind, consciousness studies, and human-compatible AI design.