My brilliant idea

René Zandbergen wrote in 2019:

Anybody who wants to present a Voynich MS solution should present the method how the text that we see in the MS was generated.
This text generation method should then explain the properties of the Voynich MS text that have been reported extensively in the past, and have been summarised to some extent in the text analysis section of this site. Most particularly, it should explain the peculiar word structure, and the variations in 'language' throughout the various sections of the MS.

I quite agree with René Zandbergen in this. And I think it was his line of thinking that inspired my brilliant idea. Or silly idea, that remains to be seen.

A concordance is an ordered list of words and names, with indications of where to find those in the book the concordance was made for. If that book is the Bible, we are talking about a Bible concordance.

Suppose a monk, or a group of monks, had a text, or were writing a text, that they wanted to be accessible only to those in the know, perhaps because they expected religious authorities of their time to be critical of it, and maybe ban it, or worse. They might use a book they had access to, and knew well, as the codebook for their look-up table cipher: the Bible. To encode their text, they used a concordance, and to speed it up, also Bible phrases they happened to know by heart. Decoding the cipher text would require an exact copy of their Bible version, and knowledge of how the monks encoded Bible locations to represent words of their text.

Which existed?

Are there any Bible concordances old enough to make this plausible? Yes, two, I learn from Wikipedia:

When, where, who, how?

1438! That’s an interesting year! It is the upper limit of the 95% probability interval for the radio-carbon dating of the VMS parchment, prepared between 1404 and 1438, meaning that that is when the animals died whose skins were used.

So now here are my conjectures, or if you will, my totally lunatic assumptions based merely on fantasy:

The Voynich Manuscript (VMS) was created by Isaac Nathan ben Kalonymus, people around him or people otherwise associated with him. They worked in or near Arles or Avignon in Southern France, not far from Northern Italy, which is consistent with the building style and plants of the drawings, and with the style of the script. Perhaps the secrecy of the text was connected with the aftermath (1378 to 1417, or 1437) of the Avignon Papacy (1309 to 1376)?

Part of the underlying text of the VMS was written in Latin, and encoded using the Vulgate Bible, and the concordance made by Hugo de Saint-Cher and his team.

Another part of the underlying text was in Hebrew. It was encoded using the Hebrew Bible and the Meïr Netib concordance, then perhaps just finished, or as a parallel co-project to demonstrate its usefulness.

In both cases the encoders had to look up every word of the underlying text, and use the concordance to find a Bible passage containing that word. Then they formed a code that indicated exactly which word from the Bible was meant, so where it is exactly. That of course is a awful lot of intensive and painstaking work. But if it is done by a team, it should be feasible. If the encoders were monks, it was literally monnikenwerk, as we say in Dutch, monks’ work. That expression doesn’t exist in English. My bilingual Van Dale dictionary translates it as drudgery or donkey work.

It wasn’t necessary to look up every word. These monks, or rabbis, or other Jewish scholars, or whoever were the creators of the VMS, were quite familiar with the Bible, has intensively studied it, and probably knew quite some passages by heart. So sometimes or often, they could skip the concordance step, and go directly to the Bible version used, where they knew the word in question occurred.

Also, for words that were frequent in the underlying language of the text, they could just randomly page through the Bible, until they happened to see the sought word. Because of it being frequent, that was likely to happen soon, and it worked with many the under­lying words, again, because they are frequent.

The fact – well, fact, just my fantasy – the fact that for what Captain Prescott H. Currier called Language A and Language B, basically the same encoding method was used, but the language of the underlying text, the codebook and look-up tables used, and details of the encoding of exact word locations in the codebook, were all different, explains why Lan­guages A and B are similar, but also notably different. That includes differing statistical properties.


More in a next episode.