Phase Transitions in Early Ontogenesis: From Pre-Architectural Organization to Cognitive Architectures

This manuscript presents a formal framework for understanding the emergence of cognitive architectures during early ontogenesis. We conceptualize early neural development as a sequence of phase transitions, where initially high-variability, pre-architectural states transform into reproducible patterns forming the first minimal cognitive architectures. The framework integrates three key components: rigid biological programs, flexible plasticity mechanisms, and structured environmental input. We introduce a quantitative measure of architectural potential Apot and a formal criterion for phase transitions based on entropy H(P) and reproducibility R(s). A minimal recurrent neural network with Hebbian plasticity illustrates the framework, showing bifurcation from exploratory dynamics to attractor states at a critical learning rate (αc). The work provides testable predictions for developmental neuroscience, including measurable changes in neural entropy, reproducibility, and region-specific timing of transitions, offering a bridge between neural dynamics and cognitive emergence. Disclosure of AI assistance This work was developed with the assistance of AI language models. The author contributed the core concepts, structure, and critical revision.