A Dynamical Model of Subjectivity

Toward an Integrative Computational Architecture for the Mind

抄録

Subjective experience unfolds continuously yet also makes discrete, qualitative jumps. To model this phenomenological duality, we propose a discrete dynamical system. This yields a control-theoretic framework governed by a Mind Topography Map (MTM) of affective gain G and cognitive bias μ. The MTM's structure is isomorphic to a cusp catastrophe. Allowing negative gain (G<0) generates error-amplifying loops that produce bistability, hysteresis, and period-doubling bifurcations. A key insight is that psychological stability is implementation-dependent. This analysis reveals a distinct zero-inertia state: the Ideal Dynamical Equilibrium (IDE). To account for the mind's complexity, we extend the model into a multidimensional framework termed Structural Gain–Bias Dynamics (SGBD). This extension resolves a fundamental computational duality of mental life: how stable psychological "states" can coexist with persistent "processes" such as rumination. We show that this computational duality emerges from the interaction matrix's symmetric (state-seeking) and skew-symmetric (process-sustaining) components. We term this principle Cognitive Phase Dynamics (CPD). To ground the framework empirically, we introduce two dimensionless indicators: DIDE to quantify the balance between inertia and responsiveness, and DCPD to distinguish states from processes. Supported by these metrics, our model serves as a computational bridge between the physics of the Free Energy Principle (FEP) and the teleology of Perceptual Control Theory (PCT). It thus offers a tractable and testable program for the future of mathematical psychology and computational psychiatry.