The Voynich Ninja

Dear Forum Users,
we have recently elaborated on a possible solution proposing lingua volgare written in shorthand. The preprint has been posted on SocArXiv.

You are not allowed to view links. Register or Login to view.

The Supplementary Information can be found here:

You are not allowed to view links. Register or Login to view.

We are looking for specialists in lingua volgare shorthand to replicate our solution and mitigate our confirmation bias. In our humble opinion, the difficulty of Voynich lies in the layered nature of the problem. We propose that the Voynich is a single substitution cipher applied to a medieval shorthand. The single substitution cipher is algorithmic and reproducible. However, the underlying text is written in shorthand and this is the real issue. Lingua volgare shorthand texts are notoriously ambiguous and even with proper language knowledge, they can remain hard to understand. The shorthand nature of the text is also why the Voynich exhibits such strange statistical properties (namely the low entropy of bigrams). Again, it is crucial to understand that a shorthand displays very different statistical properties in comparison to its source language.

We address the entropy problem in a longer explanatory text and we also identify grammatical material. We have reached out to specialists in lingua volgare shorthand personally, in order to allow for a replication by someone other than the authors (Caspari & Faccini). But of course we very much appreciate attempts by lingua volgare experts outside our immediate circle. The explanatory text can be found in the supplementary information section. We use seven criteria established by Prof. Claire Bowern to guide our argument. 

You are not allowed to view links. Register or Login to view.

Furthermore, we provide a list of 400+ words which can be read in lingua volgare. The words form part of a consistent vocabulary featuring a plethora of terms related to plants and horticulture. All words can be found in the texts of authors roughly contemporary with the creation of the Voynich. Each word in the list of the supplementary information is joined by a remark and context of the word in a source of the time. Often the example sentences feature additional words we also read from the Voynich. Here is the link to the vocabulary list (for convenience we have also supplied PDF files of the same list separated into vocabulary and comments):

You are not allowed to view links. Register or Login to view.

Our proposed solution also identifies agglutinating characteristics of the Voynich including the identification of relative pronouns and articles which we find are key to understanding the mechanisms behind the text's creation. This phenomenon finds frequent correspondences in the lingua volgare manuscripts of the time.

We look forward to constructive criticism and a well-informed discussion.

Sincerely,

Gino Caspari & Agnese Faccini

Were you able to read some longer passages or were it words and maybe short sentences? Maybe you could share some examples here?
If you have some longer texts, are they grammatical?

(02-03-2025, 01:33 PM)Rafal Wrote: You are not allowed to view links. Register or Login to view.Were you able to read some longer passages or were it words and maybe short sentences? Maybe you could share some examples here?
If you have some longer texts, are they grammatical?

See the Seven Criteria by C. Bowern document p. 3-9. You are not allowed to view links. Register or Login to view.

Downloads: 1

Was I the first to download it? Probably not, several other counters are still displaying 0 downloads for documents that I have downloaded.

Helelo Gino & Agnes,

interesting idea. As your proposed solution bases on simple substitution ("We propose that the Voynich is a single substitution cipher applied to a medieval shorthand") I'd suggest you watch the videos below which will provide some of the critical views and considerations your proposed solution will face.

You are not allowed to view links. Register or Login to view.
You are not allowed to view links. Register or Login to view.

Finally an article with references!
If I understood correctly, you determined your alphabet from two words? Your vocabulary file is not easy to consult, you added the words in EVA?

P.S. I would like to know, in German universities, do you place the methods just after the conclusions?

Quote:See the Seven Criteria by C. Bowern document p. 3-9.

These exemples are 3 short sentences. I am interested if there is something more.

Dear Scarecrow,
thank you for the input. We are aware of these statistical criticisms of a simple substitution cipher. The expectation to be able to read the text once it is transformed using our solution would be misguided. 

We have added an extensive chapter in the additional explanatory text "Seven Criteria by C. Bowern" => You are not allowed to view links. Register or Login to view. in which we explain why the entropy of bigrams is so low. It has a lot to do with the oversplitting of characters. We argue that the bigram entropy is significantly higher than the result we get if EVA is employed.

While oversplitting can account for some of the low entropy, it does not solve the problem completely (especially with regards to the word endings). Here it is absolutely crucial to understand what a medieval lingua volgare shorthand is and how it works. Shorthand scripts often disproportionately affect the endings of words. We have made an effort to explain some aspects of it based on an example in the same text. 

In short: We argue that the application of our solution to the Voynich leads to a text in linguage volgare but written in shorthand with some extreme contractions that will need additional expert interpretation. Our solution is in need of a replication by someone intimately familiar with lingua volgare and medieval shorthand texts.

(02-03-2025, 02:11 PM)Ruby Novacna Wrote: You are not allowed to view links. Register or Login to view.Finally an article with references!
If I understood correctly, you determined your alphabet from two words? Your vocabulary file is not easy to consult, you added the words in EVA?

P.S. I would like to know, in German universities, do you place the methods just after the conclusions?

Dear Ruby Novacna

The words you refer to in the main article were merely a starting hypothesis which is then further elaborated. Please refer to the main article, namely Table 1 for the transcription mechanism. We do not use EVA for transcription as we exclusively worked based on the manuscript. The oversplitting/undersplitting of EVA has its own pitfalls and a fixed transcription ultimately disregards some of the ambiguity of the original glyphs in the manuscript. Note that all 400+ words we propose in the vocabulary list are consistent with the presented solution. It therefore matters little from what words you start. Of course, there are a significant number of words that are contracted (as to be expected in a shorthand text) and these contractions introduce ambiguity. You find some of the clearest word examples in the main article.

P.S. Methods after conclusions is a format for some journals of the Springer Publishing Group.

(02-03-2025, 02:34 PM)ginocaspari Wrote: You are not allowed to view links. Register or Login to view....We do not use EVA for transcription as we exclusively worked based on the manuscript.

Using a known and available transcription is just a sign of respect to the readers, in my opinion. Like, for example, we have to write on the forum in Latin characters and in English. It does not oblige us to work in English.

(02-03-2025, 02:47 PM)Ruby Novacna Wrote: You are not allowed to view links. Register or Login to view.

(02-03-2025, 02:34 PM)ginocaspari Wrote: You are not allowed to view links. Register or Login to view....We do not use EVA for transcription as we exclusively worked based on the manuscript.

Using a known and available transcription is just a sign of respect to the readers, in my opinion. Like, for example, we have to write on the forum in Latin characters and in English. It does not oblige us to work in English.

A distinctly different transcription from both EVA and V101 is an integral part of our solution. It is of course possible to map them onto each other. However, these transcriptions (despite being a well-established go-to basis for all kinds of analyses on the Voynich) in our humble opinion are the reason why some of the statistics (e.g. entropy) are exhibiting the unusual statistical characteristics. This has nothing to do with disrespect to the readers or the immense work that has been done before on the transcriptions. We simply disagree on the separation of many of the glyphs. Ultimately, basing a solution on the actual manuscript rather than the transcriptions - which are themselves an interpretation step - is maybe more work, but it does avoid some implicit biases.

To illustrate what we mean by that, please take a look at chapter 6 in You are not allowed to view links. Register or Login to view. where we use EVA to show the changes in entropy our solution naturally brings with it.