VMs: suggestion on the author of the Voynich manuscript

[JJJRLandau@xxxxxxx]

On the Web page at URL
http://rec-puzzles.org/new/sol.pl/cryptology/Voynich
Dr. Jim Gillogly makes the following comments:

      If you want *my* guess, it's a hoax made up by [Edward
      Kelley] and an unnamed co-conspirator and sold to Rudolf
      through the reputation of John Dee (Queen Elizabeth I's
      astrologer).


      One of the most interesting Voynich discoveries was made
      by Prescott Currier, who discovered that the two different
      "hands" (visually distinct handwriting) used different
      "dialects": that is, the frequencies for pages written in
      one hand are different from those written in the other.


      My personal theory...is that it was constructed by Edward
      Kelley, John Dee's scryer, with somebody else's help (to
      explain the second handwriting) -- perhaps Dee himself,
      although he's always struck me as a credulous dupe of
      Kelley rather than a co-conspirator (cf the Angelic
      language stuff).


Having thought over the above comments, I have devised my own
theory, which is probably useful only for adding bulk to your
already-overflowing file folder of crackpot 
theories, but which I will inflict on you anyway.
     
Yes, the Voynich manuscript is a hoax and yes, Dee knew that when
he sold it to Emperor Rudolf.  Kelley was in on the plot (Dee
could not have kept secret from Kelley that he had just made a
large chunk of money) but did not play a significant role.

(I am annoyed by the theories that Kelley produced the manuscript
without Dee knowing it. Even when they were not swapping wives,
Dee and Kelley were living in such close quarters that if either
one were engaged in such a time-consuming project, the other one would
notice.)

However, Dee was only the front man in the plot.  The Voynich
manuscript was created by Queen Elizabeth's secret service chief, 
Francis Walsingham

While Dee was at Elizabeth's court, he and Walsingham knew each
other quite well.  That does not mean Walsingham liked having Dee
around.  On the contrary, he considered Dee to be a, to put it
mildly, public relations disaster who was better off anywhere
than in Elizabeth's court.

Anyway, Dee left the court in 1582.  I have no idea whether
Walsingham instigated the trip or merely took advantage of it,
but Walsingham asked Dee to be Elizabeth's secret agent while on
the Continent.  

Why?  Two reasons.  First was that Walsingham was 
placing as many agents on the Continent as he could, and Dee
might be useful.  Second, by making Dee feel that he was serving 
Elizabeth, the spy job would be an inducement to keep him out of England.

This is not to say that Dee's primary purpose for touring the
Continent was espionage.  It is just that Walsingham could always
use another agent in Europe and would not pass up the chance to
recruit Dee as a part-timer.

Whether Dee did any significant spying elsewhere is immaterial,
but eventually he arrived at Rudolf's court in Prague.  Now
Rudolf was a Hapsburg, and England was playing a dangerous game
with the Hapsburg kings of Spain (a game that would lead to the
Armada.)  It would be worth some effort to place an agent in
Rudolf's court.

Now Walsingham knew about Rudolf's obsession with the occult (one
imagines that Dee told him).  One way to get Dee into Rudolf's
inner circle was to have some apparently earth-shaking piece of
occultism for Dee to pass on.  Walsingham had no such item
available, so he did the next best thing and faked one.  The
Voynich manuscript.

Rafa T. Prinke  at URL
http://hum.amu.edu.pl/~rafalp/HERM/VMS/dee.htm claims that Dee in
Prague was strictly an outsider with little or no access to
Rudolf.  My Walsingham theory is compatible with Prinke's claim. 
Walsingham, seeing that Dee was not getting the kind of access to
the court that a secret agent required, created the Voynich
manuscript as a last-ditch ploy to provide Dee with something
that would interest Rudolf.  

As for the famous 600 ducats that Rudolf paid, Walsingham
cheerfully told Dee to keep the money.  It didn't come out of the
secret service budget and it kept Dee from handing in an expense
account.

So Walsingham put two calligraphers, an artist, and perhaps a few
other members of his staff onto the project of creating a most
authentic-looking fake manuscript.

Why two calligraphers?  Maybe to get the job done faster.  Or
maybe one broke his arm partway through and had to be replaced. 
Still another reason for having a second calligrapher will be
discussed below.

One piece of authenticity in the manuscript was the star charts,
which Dee, a leading astrologer and astronomer, could vouch for
to Rudolf.  This may be the only piece of the manuscript which
Dee helped prepare.

The artwork was to add authenticity, as in those days if a
manuscript were as valuable as Dee claimed, whoever commissioned
it would have taken the trouble to have it illuminated.  The
artwork has no relation to the text and no relevance to my
theory, but it is worth describing.

The artist had a talent for inventing non-existent flowers and
other plausible illustrations.  What about the nudes?  One
imagines the artist saying to Walsingham, "I can't do that many
unique drawings unless I go to naked broads" and Walsingham 
shrugging his shoulders and telling him to go ahead.  What about the
plumbing?  Simple.  Our artist had a bathtub fetish.  It is
comparable to a 21st-Century man who has a fetish for women's
underwear.  If he wanted to see scantily-dressed femalesd, he could 
go to a swimming pool and watch all the bikini-clad women he
wanted.  But no, he is fascinated by something secret and taboo
to males, namely female UNDERgarments.  In Elizabethan days most
women took baths only when they got caught in a heavy rain.
Hence bathtubs and their associated water pipes were virtually
secret and so rarely seen as to have an aura of taboo about them.

On to the text.  Walsingham's first decision was what alphabet to
use.  He couldn't use Greek, because nothing is written in the
Greek alphabet except Greek.  Similarly with the Cyrillic,
Arabic, and Hebrew alphabets.  That left the Latin alphabet, used
then as now for many  languages.  But in those days the Latin
alphabet was restricted to Europe and its colonies.  Now Rudolf, with
subjects that included Slavs, Austrians, Hungarians, and Italians, 
must have had the most polyglot court in Europe  On
reasonably short notice he could locate a competent translator of
any European language, including such exotica as Basque, Lapp, or
English.

(It is interesting to picture Dee being called in by Rudolf's
cryptographers to give his expert opinion on whether the text of
the manuscript were in English.)

So Walsingham had to go with either a non-European writing
system, or make up one.  Now in this era Europeans, like
Francis Xavier in China and Japan, were running across 
all sorts of strange writing systems, such as Chinese ideographs
and the Devangari alphabet of India.  Nobody in Europe at this
time could catalog all of the world's writing systems, 
so seeing an unknown alphabet would not have made Rudolf
suspicious.  There was a good chance that somebody who could read
that system would show up on the next boat from the Indies or
wherever.

This also meant that Walsingham could not use Chinese or any
other real writing system, because sooner or later Rudolf would
find someone who knew it.  Hence there was no choice but to
invent a new alphabet, which Walsingham's calligraphers did by 
plagiarizing strokes and curves from existing European alphabets. 
The kind of analysis needed to trace back these letter forms did
not yet exist in Europe, so Walsingham was safe.

Now for the text.  Walsingham realized that Rudolf would pass the
manuscript on to his Black Chamber, and knew one important fact
about cryptanalysts.  When confronted with a new cipher, they did
frequency counts.  If those did not break the cipher, they did
more elaborate frequency counts.  Hence the text had to survive
the most exacting counts that Rudolf's cryptography service could
think of.

Contrary to a statement on
http://rec-puzzles.org/new/sol.pl/cryptology/Voynich, 
it is very easy to create a random strings of letters.  Simply draw 
letters from a hat.  However, this would create an even frequency
distribution, which would shout "Viginiere!!"  The thought that
someone would encypher a lengthy text in Viginiere and THEN have
the cyphertext illuminated would be so preprosterous that even
someone as credulous as Rudolf or Senator Joseph Biden would get
suspicious.

Similar objections would occur with any attempt to create
meaningless dummy text.  Walsingham had no choice but to use a
real text in a real language and hope that Rudolf's 
Black Chamber could not identify it.

At this time there were printing presses in Spanish America,
which had produced a few books in American Indian languages that
only a handful of men in Europe could read, and none 
of these men were likely to visit Prague.  Some English traveller
or raider might have picked up a copy of a book otherwise
unavailable outside the New World, particularly in Prague.

      http://rec-puzzles.org/new/sol.pl/cryptology/Voynich reads
"According to Poundstone, a      William Bennett (see below)  has
analysed the text with a computer and finds that its entropy is
less  than any known European language, and closer to those of
Polynesian languages."

Perhaps the base language is Polynesian, but in 1580 the
Polynesians were all but unknown to Europeans and it is highly
unlikely that any text in Polynesian was available in England.  
How does the computed entropy match that of South American 
Indian languages, or Filipino languages, or those of any other newly 
discovered area in which the Spanish had set up printing presses?

I have a candidate for the book Walsingham's calligraphers used. 
It is 
     
     Confessionario para los curas de Indios : con la instrucion
     contra sus ritos ; y exhortacion para ayudar a bien morir ;
     y summa de sus privilegios ; y forma     
     de impedimentos del matrimonio / compuesto y traduzido en
     las lenguas quichua y aymara por autoridad del Concilio 
     Provincial de Lima del an~o 1583

     Publicacion:En la ciudad de los Reyes : por Antonio Ricardo
      ..., 1585

     En port. grab. xil.: anagrama de la Compania de Jesus -
     Texto del confessionario, 

     la exhortacion para ayudar a bien morir y la forma de
     impedimentos del matrimonio en castellano, quechua y aymara

As best as I can translate this into English:

     Catechism for priests to the Indians: with instruction about
     his rites, and an exhortation to die well, and a summary of
     his privileges; and the rules of marriage, composed and
     translated into the languages Quechua and Aymare by
     authorization of the Provincial Council of Lima in the year
     1583

     Published in La Ciudad de los Reyes (City of the Kings)
     by Antonio Ricardo, 1585

    Sign of the Society of Jesus.  Text of a confessional, exhortation
    for helping to die well, and forms for marriage in Spanish, Quechua,
    and Aymara

(The above citation is from www.bne.es, the on-line catalog of 
La Biblioteca Nacional de Espan~a (National Library of Spain),

I cannot identify this Ciudad de los Reyes.  An Ecuadorian
coworker says that he never heard of such a city either, but he
thinks it might be a name the Spanish applied for a while to
Cuzco, the seat of the Incan kings.

Since the book was authorized by a committee in Lima, and since
the Spanish had had printing presses in the New World since 1543,
it is most likely that this book was printed in the New World.

The New Catholic Encyclopedia identifies this book as a "Roman
Catechism" that followed the rules established by the Council of
Trent for catechisms, and states that the book was printed in
1584 rather than 1585.

We now have another reason for having two calligraphers. 
Ricardo's Catechism was printed in both Quechua and Aymara (as
well as Spanish), so one calligrapher copied from the Quechua 
text and one from the Aymara text.  Analyses of the "A" and "B"
texts show different characteristics.  These are due to differences in the
frequencies of letters and words in the two languages and also
because the two calligraphers had somewhat different rules on how
to transliterate from their respective languages into the Voynich
script.

The purpose of encyphering a document is to destroy patterns that
would aid a cryptanalyst without corrupting the plaintext. 
Walsingham had the opposite imperative.  He needed to retain the
patterns while destroying the plaintext.  Having gone to the
trouble of inventing a new alphabet, he was now faced with the
problem of a plaintext that would be easy to recover if anyone
happened to have the book it was copied from.  So he had
to do something to make the plaintext unrecoverable while
preserving most of the patterns of the plaintext.

This was not a very sophisticated age when it came to
cryptographic techniques, so Walsingham settled for a simple
technique: frequent and systematic insertion of null
characters.

I think I can identify two sets of these null characters.

I made a crude and probably not very accurate survey of the words 
in the first 31 folios (Currier transcription).  There were 4905 words of
text consisting of 1874 different words.  Of these 604 (12.3%) ended in
"OE", 370 (7.5%) ended in "8G", and 478 (9.7%) ended in "AM".  That is, 
a small number of endings appeared on a large percentage of the words used.

These statistics suggest the case-endings and other inflections of 
some European languages.  I think they are nulls, added to make Rudolf's
cryptographers feel at home, since Rudolf's people  were familiar
with German, Czech, Latin, and probably Greek, all of which have
numerous words with inflectional suffixes.

What is the rule for adding these suffixes?  A very simple one
would be to have a text in Latin or German or some similarly
inflected language next to the one which provides the
plaintext.  If the nth word of the inflected-language text had an
inflection, then this inflection would be encoded onto the end of
the nth word of the plaintext.

As for the second set of nulls, examine the copy of a page of the
Voynich manuscript that is on-line at
www.cu.jyu.fi/~paasiver/crypt/images/voy1.jpg.  I have no idea
whether this page is typical of the rest of the manuscript.

On this page there is what I call "a ponytail plant" because the
flowers resemble the clips my daughter uses to fasten her
ponytail.  Two stems of the plant run through the text and one
flower dips down into the text from above.  It is obvious that
the plant was drawn or at least blocked out on the page before
the calligrapher went to work, as the text carefully avoids
running into the drawing.  Note that the third word on the fourth
line had to slant downwards to miss the flower, and that each the
first three lines between the left-hand stem and the flower had
to be written lower than the rest of the line so as to fit
between stem and flower.

Also note that the calligrapher was at pains to try to keep a
justified right margin. On this particular page, he is also at
pains to have a smooth right margin for the parts of each line
that are interrupted by the stems of the ponytail plant.  

How does a calligrapher perform right justification?  One way is
by varying the spacing between words.  Notice that the fifth and
sixth words of the second line have an unusually small spacing
between them (so small that one might think an unusually wide
letter space had been put into a long word), whereas the spacing
between the fifth and sixth words of the third line is unusually
long.

There is another way to justify lines, if you're allowed to take
liberties with the text, and that is by inserting extra letters
(nulls) into words.  I think I can spot the null character.  It
is the letter that looks like the tower of a high-voltage power line.  
It is the first character of the first line, the one that begins with a stroke
going vertically upwards, continues on a 270 degree turn
to the left so that it emerges horizontally to the right, then
makes another 270 degree turn so that it is heading downwards,
and terminates with this second vertical stroke.

If I am reading the transcriptions properly, this is the letter that
is designated by "H" in the transcriptions.

This letter has a strange pattern of occurrences.  It is the
first character of six of the eleven lines on the page, but after
the first word of each line it occurs almost always as the second
or less often the third letter of a word.  That's a very strange
spelling convention, but it makes sense if "H" is strictly
a spacing character.  The calligrapher first wrote each line into
a grid (what oldtimers in the computer field call a "coding
form") which would tell him what was needed in the way of
spacing.  If a line looked to be a letter short, he would start
it off with "H".  As he worked his way across, if he still
found too much blank space at the end, he would insert another
"H".  Why not at the beginning of a word?  Perhaps because he
needed to write down the first letter of a word to establish how
long the space in front of the word was to be, before deciding
whether the word was long enough to justify the line or if the
line needed some additional help.

Another reason for suspecting "H"  was a spacing charcter is
that its width can be easily varied so as to match the required
spacing.  Compare how far the second letter of the fourth line is
from the left margin, compared to how much further the second
letter of the seventh line is.

http://rec-puzzles.org/new/sol.pl/cryptology/Voynich reads

     "The first and second character of Voynich words (using the
     Curva alphabet) have lower entropy than in Latin. The
     Voynich words contain more information from the third
     character onwards (in the conditional sense)  <some text
     deleted>
     "A word game to translate Latin to Voynichese must: 
          Increase predictability of word starts 
          Make words shorter 
          Maintain the length of the vocabulary."
     - Rene Zandbergen , "From digraph entropy to word entropy
        in the Voynich MS" 

If the tower is a null, then it provides a reason for the above
statements.  The entropy of the first two characters of each word
will be lowered if a particular null were frequently inserted in
those two places.  Ditto for "increase predictability of word
starts".  As for "making words shorter [than in Latin]", perhaps
the language used for plaintext already had words averaging
shorter than in Latin, or perhaps the calligrapher had a rule for
breaking long words in two to provide extra word spaces to make
right justification easier.

"A theory is only as good as the predictions it makes" so here
are a few predictions:

1.  Anybody who can write over a hundred pages of text by hand
without making the text look sloppy is a professional.  Even in
Elizabethan times there were few professional calligraphers,
since there is not that much demand for documents that have to
look pretty.  If it is possible to compare the lettering of the
Voynich Manuscript with that of manuscripts in the Latin
alphabet, then it would be possible to compare the two hands
of the Voynich MS with calligraphed documents from Elizabethan
England.  A match would not say anything about Walsingham or Dee
but would show that the MS was composed in England in or
arround Elizabeth's reign.

2.  Removing the "inflection" suffixes will make a significant
change to the statistics (it will definitely raise the entropy)
and will cause these statistics to more nearly resemble those of
a "normal" language.

3.  Removing the "H" character as being strictly for spacing
will again cause the statistics  to be more reasonable (certainly
it will lower the predictability of the first two characters of
each word).

4.  IF Walsingham kept accounts and other administrative records,
and IF these records are still extant, and IF Her Majesty's
government will allow these records to be viewed, then these
records might contain certain expenditures that could be
identified as being for Dee's manuscript.

5.  If the plaintext were from a book in a non-European language
printed by the Spanish outside Europe, it was probably a
catechism or other religious manual.  If so, then there
are some obvious probable words: "Jesus", "Christo", "santo" or
"santa", "Maria".  No matter how the rest of a Christian document
is translated, the preceding words will probably retain their
Spanish form.  Someone familiar with 16th century catechisms
could probably create a longer list of probable words.

6.  Of course if I am right about the text being the Ricardo
book, then the Voynich text can be matched against it, even
though the calligraphers are likely to have disassembled 
the book to take pages in random order and were careful not to
use matching Quechua and Aymara text.

Thank you for your patience in reading this letter.

                 Sincerely,


                James A. Landau
                systems engineer
                Basic Commerce and Industries, Inc.
                FAA Technical Center (ACB-510/BCI)
                Atlantic City International Airport NJ 08405 USA


                E-mail JJJRLandau@xxxxxxx