Encoding and Decoding Example

On the Encoding Method page I have replaced what was there with the first example of encoding and decoding.

This is Aldebaran, the label by the large star in the upper left pie slice of f68r3. Both encoding and decoding use the same letter table, with encoding taking place horizontally, and decoding vertically.

At the moment I have six other labels from f68r3 that encode and decode the same way.

There are still problems, such as there is no way of distinguishing which of the three letter value sets is used for a particular label other than trial and error. Also the problem of converting from Voynich letters to English letters. But I’m working on them.

Thoughts and comments requested.

New Page Added

I have added as a separate page above the encoding/decoding method of the VMs.

A number of people expressed difficulty understanding it, so I have laid it out in as simple and direct a fashion as I can.

Noting continued difficulty, I have modified both pages to add information that will hopefully be useful in understanding.

I realized I used an older version of the method, so updated the page and added a file showing the encoding and decoding of the word THE in the text of f68r1.

Comments and questions welcome.

Does This Look Like Voynichese?

Hoax 01

I wrote this using my “workaround” method, which lays out the text and lets the researcher change letter values at will to look for sensible text. It does not require the usage of the encoding/decoding method.

The letter values are “Set 4”, used in the paragraph text. Sets 1 – 3 are used in the labels.

Here is the text in the workaround format:

Hoax 02


I would be very interested in hearing what others think. Thank you.


This might be (read: probably is) premature, but I want to announce a possible partial translation of the text on f68r1 and r2.

I am not using my encoding/decoding method (see entry below), but the workaround I invented for my label work. By ‘workaround’ I mean a way to lay out the text and apply letter values according to the rules found by others without involving the method mechanics.

I assumed English is the language, used the Voy-101 transcription and started with two cribs on f68r1: COMET and MOON, chosen because they have letters in common and are likely to be there. I did indeed find places where they fit, and the next words to show up were OMEN and THE.
The workaround follows all of the following rules:
  • Claston’s letter frequency (i.e. Plaintext letter values are restricted to Voy letters of the same frequency level)
  • Neal’s letter substitutions and location restrictions
  • Roe’s letter order
  • Stolfi’s paradigm
  • All instances of the same Voy letter have the same plaintext value(s)

In addition, there is the rule used by the Author that is visible by the shading pattern of unused letters, that

  • in a given word, the most common value is used for all but one letter.

There are 44 words in the paragraph on f68r2. Of those 44, a valid (i.e. follows the rules) word was found for 40, or 91%. Of the remaining four, one is likely a proper name.

I have found that, generally speaking, the possibilities are limited to two or three valid words. This would explain the small vocabulary, as well as the repeated words.

For example, Line 3 word 1, EVA <tCheos>, can be either OMEGA or IMAGE, but nothing else.

EVA <or> can be DO, IN, NO, and ON, which would require four different English letters total. EVA <og> can be OF or IF.

To this point there is no indication that more than one sentence can be produced from the same Voy string. It was hard enough finding one for the two strings I have.

Rather than put the text up here, I have uploaded the two files to the MediaFire folder (link to the right) with the titles

F68r1 Paragraph Translation and F68r2 Paragraph Translation

Simplified Encoding Method

Here is a very simplified version of my VMs encoding method that gives all of the essentials of how it works. A PDF version has been added to the MediaFire folder.


The plaintext word in this example is “and”.

1. The word’s letters are moved vertically to a position directed by the letter table.
2. The letters are moved to the right and reassembled vertically in the new order.
3. The word is returned to the horizontal.
4. The plaintext letters are replaced with Voynich letters.

Encoding Method









The method is simple, elegant, and will decode as well as encode.

For some reason the text editor will not keep the formatting visible in the edit field, so I have replaced it with a jpg. Refer to the PDF for the entire entry.

Scroll down to the 14 April 2013 post to see the method using the actual letter table and a label.