Translating the Intranslatable: Traditional Chinese Medicine, Large Language Models, and the Philosophy of Medical...

The globalization of Traditional Chinese Medicine (TCM) hinges on translation practices that honor its philosophical and relational epistemologies, yet dominant approaches, including those mediated by artificial intelligence (AI), often reduce its holistic frameworks to biomedical equivalents. Terms like qi () or shen () are not merely lexical challenges; they are ontological anchors in a system where the heart (xin, ) governs emotion and moral clarity, and "wind" (feng, ) signifies cosmic disruption. When rendered reductively, whether by human translators or machine algorithms, these concepts risk becoming clinical caricatures, undermining TCM's integrity and therapeutic efficacy. This paper interrogates the epistemic tensions inherent in translating TCM, arguing that AI-driven tools like large language models (LLMs), despite their efficiency, flatten the hermeneutic and cosmological layers that define TCM as a living practice. Where AI sees "syndrome differentiation" (bianzheng, ), TCM perceives a ritualized act of reading harmony and disharmony in pulses, seasons, and bodily metaphors. Such ontological erasures compound historical asymmetries, positioning Western biomedicine as the default lens for interpreting global healing traditions. Rather than rejecting AI outright, this paper proposes a collaborative model where machine translation is guided by TCM practitioners' embodied knowledge and classical textual hermeneutics. By centering human expertise, from pulse diagnosis to the ethical dimensions of shen, translations can resist reductionism and preserve TCM's dynamic worldview. This approach transcends technocratic hegemony to foster intercultural dialogue. Ultimately, the act of translating TCM becomes a site of epistemological justice, demanding technologies that mediate, rather than master, culturally rooted ways of knowing.