The Bias Continuum Framework (BCF) for Medicine–Food Homology: A Derivable Operational Scale Across Food, Food-Medicine, and Medicine

Objective: The current Chinese catalogue of medicinal and edible substances applies a “safety–history” criterion that decides whether a substance may be admitted, but it cannot state how strong its cold–hot nature bias is after admission, where its boundaries lie relative to everyday foods and medicines, or how it should be used according to individual constitution. This study constructs an operational scoring framework that uses cold–hot nature bias as a single coordinate axis, is fully rule-disclosed and reproducibly computable, and can be extrapolated to plant-derived substances that lie outside the frozen reference tables. Methods: We propose the Bias Continuum Framework (BCF; methodological framework V1), in which foods and medicines are defined as intervals on a continuum of [−10, +10], indexed by the nature bias score B (positive = warm/hot, negative = cool/cold, zero = neutral). B is computed from three measurable dimensions—environmental habits (40%), metabolic characteristics (40%), and growth cycle (20%)—under a publicly specified operational scoring rule. Two cut-offs partition the continuum into three zones: |B|≤3, the low-bias (food-nourishment dominant) zone; 3<|B|≤6, the moderate-bias (medicine–food homology) zone; and |B|>6, the high-bias (medicine dominant) zone. The three-dimensional rules and weights were frozen and published on an international preprint platform in 2025, and no rule was retroactively changed for any substance in the present calibration. The 106 national catalogue substances were scored item by item and reported by evidence tier (L1/L2/L3a/L3b/R). Validation followed an ascending staircase: in-system consistency (60 typical foods/herbs), extrapolation of the frozen rules to 19 off-catalogue foodstuffs, and a “predict-before-check” blind prediction on 15 exotic crops and new-resource foods outside the reference tables, supplemented by a cut-off grid sensitivity analysis. Results: The 106 substances were classified as 41 low-bias (38.7%), 50 moderate-bias (47.2%), and 15 high-bias (14.2%), a unimodal distribution that bulges in the middle and tapers at both ends. All 30 everyday foods fell in the low-bias zone, and 28 of 30 recognised strongly biased herbs fell in the high-bias zone, with zero food–medicine diagonal misclassification. Of the 19 off-catalogue foodstuffs, 17 were low-bias and 2 (winter melon B=−3.2, bitter melon B=−3.3) crossed into the lower edge of the moderate-bias zone. The three-zone structure remained stable for α∈[2.5, 3.5] and β∈[5.5, 6.5]. B correlated strongly with the rank of Pharmacopoeia natures (Spearman ρ=0.910, n=106, P<0.001; ρ=0.938 in the high-confidence subset), and the plant-derived score span within a single Pharmacopoeia nature label reached about 6.9 points (the four natures govern direction rather than intensity). All 15 high-bias substances were supported by high-confidence evidence (L1/L2); a leave-one-class-out sensitivity analysis that removed all L3b and animal-source (R) entries left the high-bias membership unchanged (0% change). In the blind prediction, 13 of 15 exotic crops agreed in cold–hot direction with public reports; the two exceptions (guarana, cacao) were both rich in caffeine and theobromine (excitatory methylxanthines), precisely exposing a coverage gap of the frozen rule in the metabolic dimension. None of the 15 fell in the high-bias zone, reproducing the general pattern that everyday foods cluster in the low-bias zone. Conclusions: The BCF is a reproducible, calibratable, and extrapolatable operational scoring tool that complements rather than replaces the current catalogue: the catalogue governs legality and safety (whether a substance may be admitted or traded), whereas the continuum governs cold–hot displacement and adaptation constraints (how it should be used by constitution, dose, and frequency). The B score measures the direction and intensity of cold–hot bias and is not a direct measure of therapeutic efficacy or toxicity. Because its three-dimensional inputs are botanical measurable parameters independent of classical texts, the framework is, by design, able to compute cold–hot direction and zone membership for exotic crops and new-resource foods “for which botanical data exist but no classical textual record does”; until independent-criterion validation is completed, such results remain “values computed under the frozen rules.” The two blind-prediction reversals (the methylxanthine gap) constitute preliminary evidence of the framework’s falsifiability. Independent-scorer blind replication (inter-rater reliability) and modern pharmacological criterion association have been registered as subsequent validations.