This paper introduces Text2Tradition, a system designed to bridge the epistemological gap between modern language processing and traditional dance knowledge by translating user-generated prompts into Thai classical dance sequences. Our approach focuses on six traditional choreographic elements from No. 60 in Mae Bot Yai, a revered Thai dance repertoire, which embodies culturally specific knowledge passed down through generations. In contrast, large language models (LLMs) represent a different form of knowledge—data-driven, statistically derived, and often Western-centric. This research explores the potential of AI-mediated systems to connect traditional and contemporary art forms, highlighting the epistemological tensions and opportunities in cross-cultural translation. Text2Tradition not only preserves traditional dance forms but also fosters new interpretations and cultural co-creations, suggesting that these tensions can be harnessed to stimulate cultural dialogue and innovation.