Pharmaceutical Preparation Vocabulary and Cross-Section Connectivity in Voynich Manuscript Folio 58r

Abstract

Folio 58r of the Voynich Manuscript follows the first herbal run but is not the direct boundary to the balneological section: later herbal material and the astronomical section intervene before f75r. This folio has been variously interpreted as a pharmaceutical index, an astronomical chart, or a patient entry selector. We present a falsification-tested analysis of f58r's vocabulary class distribution and cross-section connectivity that is consistent with a pharmaceutical preparation bridge interpretation. The s-class (Honeycutt Ai Labs corridor-attested gloss s=syrupus [corridor-attested]) is one of three co-dominant line-opener classes alongside y- and d-; the lattice's high-signal cross-section anchor block for f58r lists otaly, otal, and okaiin anchors, with otal chip-annotated as routing to balneological f75v.50 and pharma f99v.27, and ary and ytar reported as multi-section cross-section operators in the line bodies. Of four prespecified falsification tests, the corridor-attestation test (T2) passes within-substrate; the class-distribution (T1), connectivity (T3), and comparator (T4) tests are reported as "consistent-with" the preparation bridge interpretation and do not constitute strict falsification of the competing patient-constitution-selector hypothesis at the token level. The internal f58r record's stored folio comparator is the Hartlieb Kräuterbuch 9 corpus at score 0.6. A separate direct structural comparison with the Cgm 7958 mansion-entry format found no repeating ordinal/name-list/entry-header parallel; neither observation is a top-match proof of function. f58r's vocabulary profile is consistent with a function specifying the pharmaceutical preparation method (syrup, decoction, soak, oil class) for plant-derived substances at the point of transition from botanical reference to balneological application. This is an internal preparation-bridge model situated within the court-physician reference-system context discussed in Honeycutt Ai Labs P-004 (DOI 10.5281/zenodo.20392093) and P-009 (DOI 10.5281/zenodo.20391967), not a sequence established by those papers.