Notes on the Voynich Manuscript - Part 12 [1992 January 30] ----------------------------------------- Concerning the Spaces in the Voynich Text Last night's wild speculation has become today's hypothesis, and it concerns the rules for inserting spaces into Voynich text. Since it is just remotely possible this hypothesis is correct, and if it is correct I don't want it named after me, I'm going to name it now: it is the Prosodic Hypothesis. In other words, the spaces do not separate words or morphemes; they separate sound clusters, syllables or feet, according to rules similar to those of metrical prose. Which hypothesis I submitted to two tests, described in this and the next note. As a working terminology, I'm going to use "words" for words, and "groups" for the things in the MS that are separated by spaces. Admonition Whether what follows is right or wrong, it has taught me a lesson I'd like to pass on. Most of us, and, I think, almost all previous would-be decipherers, have adopted, unconsciously, a modern, western approach to the MS: we have treated is as a visual object. We've been obsessed with symbols, characters, lines, words - in other words, the text as something to be read silently. We've burned our eyes over it. Et oculos habemus, sed non videmus. The mediaeval world was not like that. Their communication was much more aurally oriented; to them texts were to be read aloud. And the effective reading aloud of a text was a matter of much concern, hence the pointings, highlighting and accentuation of the text. If this MS is understandable text, it might very well have been read aloud; it might be a lectionary copy. And, just as I earlier thought the divisions might be the result of taking oral dictation, I now think they could equally well be pointings for oral recitation. Either way, I think we should focus a bit more on the text as sound, as vowel and syllable, phoneme and foot. Poetry Reconsidered In my facile dismissal of the idea that the MS is poetry, I was equally at fault. The notion of poem as text, lines neatly laid out, is, again, a modern notion. In a narrative poem on the classical model, you don't need line breaks - they can be reconstructed from the scansion. So each paragraph might indeed be a short stanza. Who, you might ask, would be so daft as to write an alchemical treatise in short verse stanzas? Well, look at 'Atalanta fugiens' by Michael Maier. The Experiment Well, on to the real stuff. The question I asked is: would a known prose text in a natural language, if divided into groups according to the rules I suspect were used in the Voynich MS, display some of the same statistical properties as that MS? Since I'd been thinking about Greek most of the day, and since the MS vowel system looked pretty Greek like, I took that as the language. I don't kave any mediaeval Greek in my library, so went back to the koine of the NT. It seemed best to take a text "at random", not selected to show off the method; given that we'd been speculating about both human languages and the angelic language of Dr Dee, one text seemed especially appropriate. Here it is, set out as prose in standard romanisation, for all us visually oriented moderns: Ean tais glo:ssais to:n anthro:po:n kai to:n angelo:n, agape:n de me: ekho:, gegona khalkoi e:kho:n e: kymbalon alalazon. kan ekho: prophe:teian kai eido: ta mysteria panta kai pasan te:n gno:sin, kan ekho: pasan te:n pistin ho:ste ore: methistanein, agape:n de me: ekho:, outhen eimi. kan pso:miso: panta ta hyparkhonta mou, kai parado: to so:ma mou, hina kaukhe:so:mai, agape:n de me ekho:, ouden o:pheloumai. he: agape: makrothumei, khre:steuetai, he: agape: ou ze:loi, ou perpereuetai, ou physioustai, ouk askhe:monei, ou ze:tei ta eaute:s, ou paroxynetai, ou logizetai to kakon, ou khairei epi te: adikia, synkhairei de te: ale:theia; panta stegei, panta pisteuei, panta elpizei, panta hypomenei. I shall now transform this according to a certain set of rules. First, I'll write the long vowels as long: eh and oh for eta and omwga. Then, I'll break the text into syllable groups where each group has a main stress at the end, either a true stress or a long vowel or diphthong. Then, I'll fuse small unstressed words onto the following group. This would be a fantastic way to read Greek, of course, but it seems to be the way Voynich text is grouped. Also, I'll kill the punctuation, and paragraph according to the traditional verses. And so, friends, here is the first sight of that long-lost mediaeval Gnostic Cathar text, the First Epistle of St Voynich to the Corinthians: Ean taisgloh ssaistohn anthroh pohn kaitohn angel ohnaga pehn demeh ekhoh gegon akhal koieh khohn ehkym balon alala zon Kan ekhoh propheh teian kaeidoh tamyste riapan takaipas antehn gnoh sinkan ekhoh pas antehn pistin hoh storeh methis tanein aga pehn demeh ekhoh outhen eimi Kanpsoh misoh pan tatahypar khontamou, kaipara dohtosoh mamou hinakau khehsoh maiaga pehn deme ekhoh oudenoh phelou mai Hehaga peh makrothumei khreh steue tai hehaga peh ouzeh loi ouperpereu etai ouphysiou stai Oukaskheh monei ouzeh tei teauteh souparoxy netai oulogi zetai tokakon Oukhai rei epiteh adikia synkhai rei deteh aleh theia Pan taste geipan tapisteu eipan telpi zeipan tahypomen ei It looks a bit familiar, doesn't it? The distribution of word length is less broad - short words have fused and long words have split. There seem to be letter correlations; the endings look grammatical but aren't, so one might be tempted to treat -oh, -eh, -ei as inflections. And I think you'd get the appearance of a lot of word repetition. Also, paragraphs would end with short groups - the odd syllables left over that can't fuse. Comments? Robert [Note: the text quoted is the famous thirteenth chapter of the First Epistle to the Corinthians; in the Westcott & Hort edition of 1893: "Though I speak with the tongues of men and of angels..."]