Research Findings

Eliminated Hypotheses

These theories were investigated in depth by specialist research agents and found to lack supporting evidence.

Supporting Findings

These investigations produced important context that feeds into the leading hypothesis.

Comparative Hypothesis Scoring

The Smoking Gun Test — Results

10 complete starred recipe paragraphs from folios f103r through f114v were decoded word-by-word using the H12 Sinhala phonetic mapping against the 9,268-entry vocabulary.

What the Decoded Paragraphs Show

Each paragraph exhibits the same structural pattern:

1. State markers open or punctuate the text — ugeda (THE dried crude drug), ugea (THE fat-soluble preparation), uteda (THE wet crude decoction)
2. Processing verbs appear in expected positions — ugala (having ground), gala (strain/filter), meda (fat/soften/knead), udena (having given)
3. Vehicles & carriers are specified — ula (spring water), ea (ghee), mea (honey), sara (essence)
4. Raw materials are named — ala (tuber/root), mula (root), upula (blue lotus), sena (senna)
5. Body targets appear in later paragraphs — ugara (throat), udara (belly/abdomen), kha (cavity/space)

Example: P107 (Folio f106v) — Strongest Pharmaceutical Reading

Reading: “[Plant]+tuber, tuber/root, having given, knead with fat, THE decoction… blue lotus… [for] belly/abdomen… THE fat-soluble preparation… ghee, make, THE infused spring-water.” — 6 LOCKED terms, 8 HIGH confidence. Contains named plant (upula = blue lotus), body target (udara = abdomen), and complete processing sequence.

Example: P083 (Folio f106r) — Disease + Processing Chain

Reading: “Bring/fetch… THE dried crude drug… [for] disease… THE dried crude drug, THE wet crude decoction… essence… [for] throat… THE fat-soluble preparation… crude base form.” — Contains disease indication (leda), body target (ugara = throat), and three distinct state markers in sequence.

Decoded Paragraph Statistics

Verdict

The decoded paragraphs do not yet read as fluent Ayurvedic instructions — too many MEDIUM-confidence glosses produce ambiguity, and the word order has not been validated against classical Sinhala pharmaceutical grammar. However, the structural signature is unmistakable:

• State markers (ugeda, ugea, uteda) appear in every paragraph — consistent with process control language
• Processing verbs (ugala, gala, meda, kara) appear in expected absolutive-chain positions
• Named plants (upula/blue lotus, sena/senna), body targets (ugara/throat, udara/abdomen), and disease terms (leda) are distributed across different paragraphs, suggesting genuine pharmacological content rather than random noise
• Compound words like ugēdala (THE-processed+having-done) and utedara (THE-decoction+rajas) follow productive morphological patterns

Full word-by-word decoding: recipe-decode-analysis.md

Ayurvedic Formula Matching

Systematic search for classical Ayurvedic formulations in the decoded vocabulary. The result reveals a pharmaceutical system rather than a named formulary.

Classical Drug Vehicles — Complete Set Confirmed

All four classical Ayurvedic drug vehicles (Chaturbhaga) present. This is the foundation of Bhaishajya Kalpana.

Classical Dosage Forms — 5 of 10 Present

Named Formula Search

Strongest Confirmed Compounds

Interpretation

The vocabulary describes a pharmaceutical processing system, not a formulary of named preparations. The complete infrastructure is present (all vehicles, 5 dosage forms, grinding/straining/filtering verbs, state markers) but the full pharmacopoeia of named plants is sparse. This is consistent with a manufacturing manual that assumes the practitioner already knows the ingredients — what the manual teaches is how to process them.

Full analysis: formula-matching-analysis.md

Recipe Section Visual Markup: The Star Encoding System

The recipe section (f103r–f116v) contains ~345–347 starred paragraphs across 23 surviving pages. These stars are not simple paragraph markers — they constitute a three-layer visual encoding system.

Three Independent Encoding Layers

Minimum 8 possible star states (2×2×2), up to 12 if paint has 3 levels. Far too much combinatorial space for decoration.

Key Anomalies

• f106r companion star: One paragraph marked with a double star — the only instance in the entire section. Implies semantic distinction (cross-reference? compound recipe?)
• f103r forgery: 19 tail-less stars crammed into margin after the section was written. The forger couldn’t replicate the tails — proving tails carry meaning
• ~359 total entries: Matches Jean le Bègue’s recipe compilation (MS. 6741, Paris, 1431) with 359 itemized recipes

Hypothesis Verdicts

Integration with Manufacturing Hypothesis

If the ~359 starred entries are SOPs for a pharmaceutical enterprise, the three-layer star system becomes a classification and cross-referencing index: dark stars = entries with pharmaceutical section counterparts, center dots = herbal section cross-references, red paint = dangerous/toxic ingredients (consistent with Agada Tantra content).

Full analysis: star-symbol-analysis.md

Pharmaceutical Section: Raw Material Specification Sheets

The pharmaceutical section (f87r–f102v) contains 57 rows of plant fragments (~239 individual drawings), each row paired with a labeled container. These are not whole plants — they show which specific parts to use.

The Visual Layout Pattern

D’Imperio: “a jar…at the left margin of each such row, irresistibly suggesting that the plants in that row were used to make up the compound prescription symbolised by that jar.”

Two-Stage Identification System

Scholars confirm pharmaceutical drawings are copies of portions of herbal section plants (Th. Petersen, Mark Knowles). Confirmed match: f37v (herbal) = item 203 on f102r1 (pharmaceutical).

Ayurvedic Principle: Plant Part Determines Preparation

In the Panchavidha Kashaya Kalpana (Five Basic Preparations), the plant part dictates the processing method:

This is exactly what the pharmaceutical section encodes visually: plant part next to container = which material goes into which preparation method.

Container Typology Matches Italian Conventions

Two main container forms: simple cylindrical (Quire 19) and elaborate ornate (Quire 15), progressing from basic to complex. Container shapes conform to standardized 15th-century Italian apothecary jar iconography where shape encodes preparation type. Cross-section root illustrations next to containers are a device found specifically in Italian pharmacy manuscripts from the Veneto region.

Verdict

WELL-SUPPORTED — The pharmaceutical section functions as raw material specification sheets: visual SOPs linking specific plant parts to specific preparation methods via labeled container pairings. The visual layout, H12 state-marker vocabulary (K-dry / T-wet pathways), herbal section cross-references, and Italian pharmaceutical manuscript conventions all converge.

Full analysis: plant-parts-analysis.md

Grammatical Coherence: Does It Read as Sinhala?

A specialist analysis evaluated whether the 10 decoded recipe paragraphs exhibit valid classical Sinhala pharmaceutical sentence structure. The result: 5/10 overall — partially consistent, with compound morphology as the strongest signal.

Dimension-by-Dimension Scores

What Works

• Pharmaceutical vocabulary categories are correct: ingredients, vehicles, process verbs, state markers, dosage forms, disease terms, anatomical terms
• Compound words follow valid Sinhala formation rules (Tappurusa determinative compounds)
• Absolutive participles (-la suffix) are present and occasionally form valid chains: ugala → gala → tamea (having ground → strain → for honey)
• Best paragraph (P107, f106v) scores 6/10: blue lotus + abdomen + fat processing + ghee + make — with a clear final verb (kara/make)

What Doesn’t Work

• State markers dominate every paragraph (4:1 ratio over process verbs) — authentic texts use them as labels, not refrains
• No measurement units found (pala, karsha, tola, masha) — pharmaceutical instructions without dosages are incomplete
• The “u-” definite prefix proposed for state markers (ugeda = “THE dried drug”) is not attested in classical Sinhala grammar
• Clause boundaries are unclear — long nominal sequences without verb intervention

Three Possible Explanations

1. Real Sinhala, partial decoding: The phonetic mapping captures genuine vocabulary but introduces word-boundary errors, disrupting clause structure. The underlying text may be authentic with higher grammatical coherence than the current decoding reveals.
2. Catalog, not recipes: The text may be a pharmaceutical index or reference table listing preparations and their states, rather than step-by-step instructions. This explains state-marker dominance and absent measurements.
3. Statistical coincidence: The H12 mapping produces CV syllables that overlap with common Sinhala words by chance, without the underlying text being Sinhala. The weak clause structure reflects random phonetic output.

Full analysis: ayurvedic-grammar-patterns.md

Full Research Reports

23 research documents produced by specialist expert agents, totalling approximately 400,000 words of analysis.

Additional earlier reports: paleo-comparison.md, linguistic-comparison.md, historical-transmission.md, voynich-asian-scholarship.md, voynich-glyph-scholarship.md, sinhala-script-comparison.md, khojki-deep-dive.md, visual-similarity-survey.md, sign-language-hypothesis.md