A phonosemantic grounding framework that uses Sanskrit articulatory phonology to create physically grounded AI embeddings. Includes clustering experiments on 150 Sanskrit verbal roots and a 10-dimensional coordinate system derived from speech production anatomy.
natural-language-processing representation-learning panini state-space-models interpretable-ai embodied-cognition sanskrit-grammar inductive-bias semantic-grounding phonosemantics source-filter-theory articulatory-phonology
-
Updated
Apr 19, 2026 - Jupyter Notebook