FarnetAI

This research group endeavors to harness the analytical power of large language models through the theoretical lens of cultural linguistics, aiming to decode how cultural artifacts drive semantic evolution within specific societies. By training these models on temporally-layered corpora—spanning literary texts, social media discourse, broadcast media, and digital archives—the project seeks to trace the subtle reconfiguration of conceptual categories and the shifting connotations of culturally-loaded terms. A paradigmatic case study involves tracking how a widely-circulated protest anthem or viral musical composition can progressively alter the public's collective understanding of foundational concepts like "father," transforming it from a purely biological or patriarchal signifier into a symbol of either oppressive authority or revolutionary sacrifice, depending on the ideological context and lyrical narrative. Through computational analysis of collocation patterns, sentiment trajectories, and metaphorical extensions across millions of data points, the methodology reveals the precise mechanisms by which artistic expression infiltrates everyday language, renegotiates social values, and ultimately redefines the semantic boundaries of kinship terminology in ways that resist traditional lexicographic capture but become unmistakably visible in the vector spaces of culturally-attuned AI models.