A Nonlinear Dynamical Process Model of Semantic Priming: RT Tail Acceleration on Discrete Semantic Graphs

Nonlinear field-theoretic model of “meaning activation and conscious selection” on a semantic graph. A meaning manifold is constructed from fixed sentence embeddings (sentence-transformers/all-MiniLM-L6-v2, 384-d). Dynamics are modeled by a focusing nonlinear Schrödinger / Gross–Pitaevskii equation on a k-NN graph with an external potential derived from an empirical probability measure μ (frequency/salience). The selection time T* is defined operationally as the first hitting time of a peak-density threshold; a minimal observation model links it to reaction time as RT = T* + ε. The main testable prediction is a nonlinear acceleration of the right tail of RT distributions (vincentile regression) under high-arousal prime × ultra-short SOA conditions.Files: PDF (main text), DOCX (editable), README (structure and citation).AI assistance: AI language models (Anthropic Claude and xAI Grok) were used in a limited way for editorial support and organizational suggestions. The author takes full responsibility for the scientific content and any remaining errors. v4 (2026-03-07): This version has been substantially revised relative to earlier releases. The title has been updated to reflect the revised scope and framing. Main changes include updated figures, revised text structure, and refreshed supplementary materials. Please cite the version-specific DOI for this release.