A Candidate Decipherment of the Voynich Manuscript: Evidence for a Spoken Elu-Sinhala Pharmaceutical Register (V5)

Version 5 of the candidate decipherment of the Voynich Manuscript, identifying it as a 15th-century Sri Lankan Elu-Sinhala pharmaceutical text. This version supersedes V4 (2026-05-09). Major new findings in V5: Rival-language tournament expanded to 27 control corpora across 11 tradition families — no tested rival beats Sri Lankan medical controls on the repeated locked-anchor metric (Sārārtha Saṃgrahaya 66.67% vs Unani/Tamil/Siddha/European 0–15%); BALNEO section recharacterised as sneha-karma (medicated oil preparation) section based on dominant vocabulary (q-keda, meda, ch-ea, gala); six manuscript sections identified as six functional components of an Ayurvedic pharmacopoeia/formulary system; pesaca (piśāca = evil spirit/goblin), sole occurrence in 36,633 tokens, opens f113r as the strongest current text anchor for a dedicated bhūta-cikitsā chapter; senna (Cassia angustifolia) confirmed as pharmacological backbone — 145 tokens / 84 folios under the fixed-gloss rule, present in all six major sections; iron-eye treatment cluster documented as three-section cross-reference (HERBAL f55r → ASTRO f73r → RECIPE f116r) with no shared folio; canonical direct-token plant identification table established (T1/T2 criteria: token + visual + cross-section recurrence); Team B full validation suite: 16/16 steps passed on frozen V17/V21 decoder (DB SHA256: 5c8b23722e395b8454ce56a53879ac91f6ea398753f76a0184a9c562228d5385). P(wrong) ~7-10%, unchanged from V4. Corpus gaps acknowledged: Arabic/Persian aqrabadhin prose, Tibetan Sowa Rigpa formulary text, and historical Tamil/Siddha and Malayalam prose comparators remain unacquired; a definitive rival-family closure claim is not made pending these acquisitions. Licensing: Original content CC-BY-4.0. Bundled third-party corpora retain their own licenses; see LICENSE.