Voynich Reconsidered: the third name?

On the first page of the Voynich manuscript (folio f1r), there are three short strings of glyphs, in which one might suspect the representation of a proper name. In previous articles on this platform, I examined the first and second of these strings.

The third string, with its keyboard assignments in Glen Claston’s v101 transliteration, is as follows:


A right-justified sequence of glyphs on page f1r, line 21 of the Voynich manuscript. Author’s analysis. Image credit: Beinecke Rare Book and Manuscript Library.

This sequence of glyphs is unique. It occurs nowhere else in the Voynich manuscript. We might therefore conjecture that it represents a proper name.

To my mind, the main uncertainty of interpretation is in the second “word”, in which there is a string that looks like {iii}, and what appears to be a final {N}. However, the string {8aiiiN} does not figure as a “word” in the Voynich manuscript; neither do the alternatives {8aim}, {8aiin}, {8aIn}, {8aIiN} or {8aiIN} On the basis of “word” counts, the most probable reading appears to be {okoe 8aM}.

On this reading, both “words” exist elsewhere in the Voynich manuscript. Both conform to the rules of sequencing glyphs within “words” which Massimiliano Zattera has called the “slot alphabet”.


Voynich manuscript, page f1r, line 21: three interpretations of the glyphs, and the permitted "slots” in Zattera’s “slot alphabet”. Author’s analysis.

Mappings

As with the first and second glyph sequences on page f1r, I attempted mappings of the sequence {okoe 8aM} from various alternative transliterations of the Voynich manuscript, to medieval Italian. My corpus of reference is OVI (Opera del Vocabolario Italiano).

In all of these exercises, I could see great uncertainty in the mapping of the glyph {M}, which (if it is a single glyph) occurs only 45 times in the whole of the v101 transliteration. Therefore, if it maps to any letter in Italian, it must be a relatively rare letter.

Below are selected mappings that, to my mind, seemed more plausible than others:


Selected transliterations of the Voynich manuscript, and corresponding mappings of the glyph string {okoe 8aM} to text strings in medieval Italian. RSQ = correlation between glyph frequencies and letter frequencies. AFD = average absolute difference in frequencies between glyphs and equally-ranked letters. Author’s analysis.

As far as the OVI corpus shows, only the text strings NOY and NOJ (both archaic forms of "we") are words in medieval Italian.

As with the first and second strings on page f1r, I can imagine other possible mappings, including the following:

• mappings based on other transliterations: for example, with the initial {o} treated as a wild card, or as an abbreviation sign
• a flexible mapping of the glyph {M), which has the right frequency to map to any of the Italian letters É, Ù, X, or K, or (less probably) the rare Y, J, Ó or W
• or a re-interpretation of the glyph {M} as {im} or {iin} or {iiiN}; and in turn, we can create permutations with {ii} re-interpreted as {I}
• re-ordering of the letters within the text strings: for example, we could re-order EDES as SEDE (“seats”, or “sees” in the sense of seats of the church); and NOK as KON (archaic "with")
• and of course, mappings based on other languages than Italian.