Design research supporting practitioners' and researchers' reasoning through multiple perspectives on classical knowledge
This research and development project structures classical East Asian medical texts, abdominal diagnosis, acupuncture information, and materia medica knowledge so that they can be consulted, compared, and recorded within MLMN's multilayer framework. It does not automate diagnosis or prescribing through AI. It is designed as an environment for practitioners and researchers to organize information from several perspectives and broaden their reasoning.
East Asian clinical medicine has been supported by long experience and accumulated classical knowledge. Contemporary practitioners must nevertheless handle much information in limited time, and cross-referencing classical literature, abdominal findings, point knowledge, and distinctions between medicinals is not easy. The system aims to reduce this reference burden and support the organization of clinical reasoning.
Practitioners themselves decide diagnoses and prescriptions. The system supports comparison, organization, and structuring of information before decisions are made; it does not replace those decisions.
Before placing information across the layers, this project explicitly records what was obtained from which examination. The five core inputs are abdominal examination, back-shu examination, tongue diagnosis, pulse diagnosis, and interview. They are supplemented by the time course from onset to the present and records of what changed after treatment.
| Source | Main Original Information Recorded |
|---|---|
| Abdominal Diagnosis | Tension, softness, resistance, tenderness, temperature, pulsation, regional and left–right differences |
| Back-Shu Examination | Tension, depression, tenderness, temperature, skin, point responses, and left–right differences around back-shu points |
| Tongue Diagnosis | Tongue body and shape, moisture/dryness, cracks, tooth marks, and the color, thickness, and distribution of coating |
| Pulse Diagnosis | Position, superficial/deep, slow/rapid, strong/weak, large/small, long/short, tension, slippery/choppy quality, and left–right differences |
| Interview | Chief complaint, onset, preceding events, aggravation and relief, history, lifestyle, environment, and previous treatment responses |
The five inputs are not immediately compressed into one pattern: original observations and candidate interpretations are separated. Nor is any input fixed to a single layer. Information is placed across the layers while distinguishing convergence supporting a hypothesis, discrepancies challenging it, and cases that cannot yet be judged.
Traditional Kampo questionnaires have largely used static checklists of separate symptoms. This project examines the following approaches to interview structure:
These remain design proposals. Implementation and operation require considerable further validation, including the intended practitioner group, input burden, data management, and safety.
Preparation status differs by text within the corpus referenced by the system.
Structured holdings include 80 Suwen chapters, 81 Lingshu chapters, 81 Nanjing difficulties, 728 Shanghan Lun passages, 1,110 Jingui Yaolue passages, Shennong Bencao Jing, the introductory sections of Bencao Gangmu, Gezhi Yulun, Neiwaishang Bianhuo Lun, 306 pages of Igaku Keigo, Mondosaku, Shanghan Zabing Lun Leibian, and Sentetsu Iwa. For each text's verification status, see Safety & Disclaimer and the individual pages.
Open the Free 24-Page Viewer →
The full 306-page edition is being prepared for members. Please contact us if you wish to view it.
The project uses AI for OCR of classical literature, text normalization, and search-index construction. These are primarily assistance technologies in the preparation process, not the agents of clinical judgment. Our consistent principle is that AI supports information processing rather than determining diagnoses or prescriptions.
The system remains under continuing research and development, with functions and content added or revised as needed. The timing, scope, and conditions for offering a clinically usable system are not yet fixed. Preparation will proceed in stages as research advances.