Silk has long served as both a material and a medium of cultural expression in Chinese civilization. During the Tang (618-907 AD) and Song (960-1279 AD) dynasties, the interplay between silk textures and chromatic symbolism reflected not only technological advances in textile production but also the aesthetic and philosophical values of the time. This study applies computational linguistic techniques to model the semantic relationships between silk-related terms and color descriptors within a large corpus of classical Chinese texts. Using a Word2Vec neural network trained on the Complete Tang Poems and Complete Song Ci-poems, the research constructs a high-dimensional semantic space to quantify associations between textile types (e.g., brocade, gauze, plain silk) and color attributes. Quantitative analysis reveals that red silk threads are closely linked with concepts of affection, destiny, and emotional tension, while white plain silk corresponds to purity, sorrow, and solitude. These associations mirror the dual symbolic system of warmth and coldness embedded in traditional Chinese textile culture. By combining computational modeling with cultural interpretation, this research demonstrates how digital methods can illuminate the intangible dimensions of textile heritage. The study contributes to the broader field of textile and cultural heritage studies by providing a replicable framework for integrating artificial intelligence with historical material analysis, thereby advancing digital preservation and interpretation of silk culture.

silk color heritage, Tang and Song China, textile culture, semantic modeling, Word2Vec