This page contains characters in the “EVA” font.  You may obtain the EVA font at http://web.bham.ac.uk/G.Landini/evmt/eva.htm

 

Various encryption schemes that may have been used in encoding the Voynich manuscript

 

Any one of the following methods (or a mixture of methods) of encryption may have been used in encoding the Voynich manuscript:

 

Ancient Character Replacement (I)

In this scheme we assume that all recognizable characters are just what we think they there.  That is, Voynich “a” converts to character “a”, Voynich “o” is character “o”, etc.  Other unusual characters are from some other (possibly unknown) character scheme.  This scheme also assumes that the text is read exactly as written.  For instance, assuming that “ora” is “ora”, this might be “gold” in Spanish.

 

Most of the Voynich characters, while appearing strange to modern readers, are actually familiar in other languages over time.  For instance, the character “8” has been used in various forms over time for the letters f in Italic (Italy) and Etruscan alphabets, m in some Roman script, the “oui” sound in 16th and 17th century French, as well as appearing as a character in proto-Sinaitic and proto-Phoenician languages.

 

Attempts at this kind of translation have been ineffective for the Voynich Manuscript to date.  This may indicate that a different form of encryption was used, or that the language beneath the characters is invented or presently unknown.

 

Ancient Character Replacement (II)

This scheme is the same as mentioned above, but the underlying language has been “abbreviated” or is “vowel-less”. 

 

There have been examples of languages over time where vowels were not incorporated into the written language.  The Dead Sea scrolls, for instance, are written in a vowel-less version of Hebrew.  While vowels are greatly helpful in ensuring that the written text is read as intended, in many cases they are unnecessary.  For instance, “cn y rd ths txt?” could be “can you read this text”. 

 

Abbreviated text is text written with some characters that would be obvious to the reader omitted.  This is seen in many written documents that are specific to some area of interest where the reader would be able to easily understand the gist of the statement without every detail spelled out.  For example, in a French military document one might find “le capt du Grenr coe et la pre coe ..” where capitaine is shortened to capt, Grenadier is shortened to grenr, compagnie is coe and Premiere is pre. 

 

Various claims have been made for solutions of the Voynich translation in this kind of translation, including John Stojko who claims the manuscript is in vowel-less Ukrainian (web site http://home.att.net/%7Eoko/home.htm)  These claims have not been widely accepted.

 

Ancient Character Replacement (III)

The text may be in a simple replacement scheme, but still be difficult to translate due to randomized spacing.  When trying to convert text into modern characters we tend to try and look at words that we assume would be placed in the text.  For instance, we may see an astronomical picture and assume the word “star” appears in the text.  If the text was written with randomized spacing we may be substituting the wrong characters into the wrong text.

 

For instance, if I wrote “the star in Taurus is covered by the moon on Tuesday” with randomized spacing we may see “th est arinTau rusis c over edby th emoo on on tue s day”  note that if we were looking for the word “star” we might begin by finding 4 character words and start the replacement.  However, in the text above the 4 letter words are “over”, “edby”, and “emoo”.  In each of these cases replacement of characters (ie s for e, t for d, a for b and r for y) one-for-one would produce gibberish.  

 

Ancient Character Replacement (IV)

Again, we assume that the characters are easily convertible to modern equivalents,  but in this method spurious characters are introduced to make translation more difficult.  A simple example of this may be to insert the letter “z” in random places throughout the text: “czazzn yzou rzeadz ztzhis tzzezxzzt?”  By skipping the “z” character the text becomes readable.  Adding more spurious characters makes translation even more difficult as translation attempts tend to produce nonsensical words if you look at them as single units.

 

Simple Character Replacement

This strategy is simply replacing one Voynich character for another character used in another language.   The characters that are familiar to us are not necessarily what they seem to be. 

 

For instance, once could put the Voynich characters into a substitution grid and just start replacing characters.  An example of this strategy may be as follows:

 

Voynich  Modern

Q         s        

o                    o

t                   st

a                    a

l                     i

iin        ud

k                    lt

e                    e

8                    m

y                    y

c                    r

 

Voynich Text:               qotaiin qokaiin okedy l kchdy

Modern Text:               sostaud     soltaud    oltemy  i  ltremy

           

The above was not a serious attempt at trying to make the translation come out to resemble a known language, but all attempts at this kind of replacement strategy has resulted in so much gibberish.  While key words can be found, the surrounding text is usually nonsensical.

 

Simple Character Replacement II, III, IV

Schemes as found in the Ancient Character Replacement strategies II-IV (vowel-less, randomized spacing, nonsense characters) could also be introduced into this schema making the translation more difficult.

 

Advanced Character Replacement Strategies

 

There are many methods where more advanced character replacement strategies could be employed.  A few examples follow, but this is by no means an exhaustive list.

 

Character Position Replacement

A different key could be used for each character position in the “word”.  For instance, in the first position “o” could represent “t”, but in the second position “o” could represent “r”, etc.  thus “oo” could be “tr”

One way to do this is to create a series of alphabets (let’s create a 3-key example)

 

 

0 abcdefghijklmnopqrstuvwxyz

   ======================

1 bcdefghijklmnopqrstuvwxyza

            2 cdefghijklmnopqrstuvwxyzab

            3 defghijklmnopqrstuvwxyzabc

 

             So a message such as “go left” would be decoded as follows:

            “g” is the first letter so we follow “g” down from the 0 row to the 1 row and get “h”.

            “o” is the second letter so we follow “o” down from the 0 row to the 2 row and get “q”

            when we reach the 4th letter we go back to row 1 and so on.

            In the end we get “hq ofhw”  Note that “h” appears twice – but actually represents g in the first instance and f in the second.

 

It is unlikely that this method was used in Voynich.  Too many words begin with “o”.  In this method all these words would still begin with  the same letter, which is unlikely – unless the underlying language had a multitude of words that began with a single letter.

 

An alternative to this would be to supply just the numbers for the above list.  For instance, if I provided 12213 you could read the first letter in 1, then the second letter in 2 then the third letter in 2, then the fourth letter in 3…etc.  Even if the Voynich characters represented numbers we would have the same problem as noted above.  The first letter in Voynich script often is “o”.  If that represented a 1, for instance, you would still decode this by the first letter on the 1 line.  Which would be the same letter each time.

 

A final alternative would me to write the Voynich “numbers” in line across such a scheme, then read down.

 

Example:

 

Assume Voynich numbers are as follows: o=1 r=2 y=3 i=4 l=5 k=6 s=7 d=8 ch=9 

 

Ror y ory or

212 3 123 12

abcdefghijkl..

bcdefghijklm..

cdefghijklmn..

etc.

 

In this case many words could begin with “o” (1) and still not represent the same letter.

 

Key Replacement

            A variation on the above is to retain the list of codes (1-3 above) and add a key.

 

For example, the encoder could choose a sentence that would be known to the recipient.  The sentence is written down and the text to be encoded is placed beneath.

 

     This is the key sentence

     Some se cre tme ssage

 

Instead of cycling 1,2,3 as in the example above we now encode the text by looking at the cipher to see which row we should read from. (for s we look for the row that begins with a “t”, for the “o” the row that begins with h, etc.  The advantage to this kind of cipher is that the list of 26 rows for the encoding can be well known.  You can even publicize it!  Without knowing the key sentence for which to do the decryption it does no good.

 

List more strategies here

 

Numerical Replacement (I)

In this scheme a key sentence is constructed and a number is used to refer to the position of the letter in the sentence.

For instance, using a key of “Thequickbrowndogjumpedoverthelazytranslator” once may translate “can you read this” as

7 31 13  33 15 5  10 3 31 14  1 2 6 38

Each Voynich character could refer to a number, rather than being a character.  In this case the translator would need to know a) the numbering scheme and b) the key used.  If the text is meant to be generally readable by those “in the know” both a and b may be included in the text.  If the audience for the document was small, and connected in some way the translation scheme and key may have been passed orally.  In that case the translation of the document would be extremely difficult.

 

List more strategies here

 

 

Comments?  Email: [email protected]