back to J.VS front-page

Archive of communications of the Journal Of Voynich Studies

Index of Subjects in all Volumes

Vol. IV, 2010

From: Berj N. Ensanian KI3U
To: Journal of Voynich Studies
Sent: 1-4-2010 8:50:12 PM EST

J.VS: Mathematical Sparse Coding analysis of illustrative artworks

Dear Colleagues

Here in a 4 JAN 2010 BBC online news article is some interesting information on mathematical sparse coding techniques
as used in the analysis of image art, specifically to distinguish genuines from fakes:

" Computer method 'spots art fakes'

A simple method to distinguish artistic fakes and imitations has been demonstrated by researchers. The approach, known as
"sparse coding", builds a virtual library of an artist's works and breaks them down into the simplest possible visual elements.
Verifiable works by that artist can be rebuilt using varying proportions of those simple elements, while imitators' works cannot.
The work is reported in Proceedings of the National Academy of Sciences. "

" The method works by dividing digital versions of all of an artist's confirmed works into 144 squares - 12 columns of 12 rows each.
Then a set of "basis functions" is constructed - initially a set of random shapes and forms in black and white. A computer then
modifies them until, for any given cut-down piece of the artist's work, some subset of the basis functions can be combined in some
proportion to recreate the piece. The basis functions are refined further to ensure that the smallest possible number of them is required
to generate any given piece - they are the "sparsest" set of functions that reproduces the artist's work. "

" However, Professor Rockmore said that although authentication of works was an application that would appeal to many people,
sparse coding could lend its analysis to a number of problems in the study of art. "Our hope is that it becomes more of what people
call technical art history," he told BBC News. "

Very good. The illustrations in the Voynich manuscript would seem to be great candidates for this kind of analysis. Check them to see
not only if the one same illustrator was at work throughout the VMS, but also of course to have a data set which can be compared with
non-VMS illustrations.

One wonders if an advanced version of this technique could be fruitful with the Voynich text - treating the handscript as if it were an illustration.
Of course here I am thinking most curiously about the hypothetical steganographic hand-script text-art in VMS folios, in particular f76r,
where in the upper half of that page's text I have proposed is embedded the illustration of a man, perhaps the Voynich manuscript author's portrait,
holding with both hands to his eyes a rectangular optical filter plate, presumably indicating that sophisticated optical techniques are required in order
to see what he has hidden in his mysterious manuscript. [1]

Berj / KI3U

[1] Various pictures of the hypothetical optical physicist and presumed VMS-author, embedded as steganographic handscript text-art
in Voynich folio f76r, are deposited in the J.VS Library:


are suitable for blinking, and are available in J.VS Lib. deposit # 18-1-2008-04-15 :

Several other pictures, including more blinks, are in J.VS Lib. deposit # 14-1-2007-10-22 :

Extensive discussion, with earlier references, of this f76r curiosity (as well as others) is found in J.VS communication # 156
(and Re:'s # 157, # 163), "J.VS: The Voynich hypothetical f76r text-art portrait: how was it done?" (Vol. II, 4 FEB 2008):

Further discussions are found in J.VS Volume II communications # 158, 161, 164, 181, 182, 183, 185, 191, and
Volume III communication # 317:

From: Berj N. Ensanian KI3U
To: Journal of Voynich Studies
Sent: 1-11-2010 11:29:21 PM EST

J.VS: J. Malcom Bird's June 1921 VMS article in Scientific American

Dear Colleagues

J. Malcom Bird's June 1921 article on the Voynich Manuscript in the journal Scientific American:

The Roger Bacon Manuscript
Investigations Into Its History, and the Efforts to Decipher It

is now available online via google Books:

Berj / KI3U

From: Berj N. Ensanian KI3U
To: Journal of Voynich Studies
Sent: 1-19-2010 9:29:21 PM EST

J.VS: The phonetisation of the Voynich Manuscript text: Voynich text to synthetic speech

Dear Colleagues

As you recall, a couple of years ago our colleague Jan Hurych began urging experiments with phoneticizing VMS text
by means of speech synthesis, that is coverting a Voynich text transcript into readable/listenable voice via a conversion table
and a voice program.

Jan last night let me hear some of his recent results, and I was very excited by them and urged him to let others hear this also,
even though he says he is not yet ready to write up a paper on this work. However Jan this morning did put one sample
of his experimentation online, a sample of Voynich f2v:


More will follow soon.

Berj / KI3U

From: Berj N. Ensanian KI3U
To: Journal of Voynich Studies
Sent: 1-22-2010 1:28:02 PM EST

J.VS: Slime Mold Networks and the VMS

Dear Colleagues

Here is a BBC news online article dated 22 JAN 2010 :

" Engineers 'can learn from slime' "

" The way fungus-like slime moulds grow could help engineers design wireless communication networks.
Scientists drew this conclusion after observing a slime mould as it grew into a network that was
almost identical to the Tokyo rail system. The scientists describe their ideas for "biologically
inspired networks" in the journal Science. They have incorporated the slime mould's efficient
strategy into a mathematical formula. This "slime formula" could help engineers develop better, more
efficient designs. "

" One of the researchers, Dr Mark Fricker from Oxford University, UK, told BBC News that the whole
idea of using slime moulds in this way came from Toshiyuki Nakagaki, a scientist also based at
Hokkaido University. A decade ago, Dr Nakagaki showed that the slime could find the most efficient
way through a maze. "

So then what does this have to do with the Voynich Manuscript?

Well, first of all, botanical concepts being a major component of the VMS, it is entirely
conceivable that the VMS author concerned him/herself to some degree with slime mold. Indeed our
colleague Dana Scott, who has long been compiling plausible VMS botanical identifications, years ago
brought up the possibility of slime, at least green slime, being represented in VMS illustrations.

This idea becomes even more interesting when we look at pictures of slime molds in the classic 1904
work of the polymath scientist Ernst Haeckel: Kunstformen der Natur (Artforms of Nature).
Haeckel's work is in general a valuable resource for investigation of the VMS illustrations.

In his book, specifically concerning slime molds, Haeckel's illustration of Mycetozoa :

gets up my VMS interest - compare for example Voynich illustration f83v, the part where two VMS
ladies / nymphs are each holding green bulbs excreting a stream of something, with the Haeckel
specimen at the right of bottom-center, which shows excretions emanating from soccer-ball like
bulbs. This also harkens back to the possibility, actually I myself would say near certainity, that
the VMS author employed optical magnifications during work, perhaps even with early microscopes.

Needless to say, slime can be symbolic, and within old suspicions that the VMS botanicals often
suggest shematics of some kind rather than straightforward if crude plants, we might fathom f83v
using slime symbolically in some way.

Moreover, if we are correct in this tentative f83v interpretation, then on account of similar
graphics in other VMS folios, even in non-botanical sections of the VMS, for example f86r, and
certainly the dramatic nine-rosettes foldout, we can then entertain that perhaps the VMS author in
some way conceived of networks of slime. Perhaps there are to be recognized among the strange VMS
illustrations some slime networks providing pointers to navigating the VMS maze.

Hence, taking in the later stages of the developing slime network shown in the video-clip of the
above BBC article, compare with the general organization of the Voynich nine-rosettes foldout.
There may be more slime and slime networks in the greater Voynich picture than we generally appreciate.

Berj / KI3U

From: Berj N. Ensanian KI3U
To: Journal of Voynich Studies
Sent: 1-24-2010 7:08:52 PM EST

J.VS: Petrus Barschius of Cracow; 1647

FYI from Berj / KI3U and Greg Stachowski (mostly Greg's work) :

Google books has scanned and made available for download the following,

“Statuta nec non Liber Promotionum Philosophorum Ordinis in Universitate Studiorum Jagiellonica ab
anno 1402 ad annum 1849.”

which appears to be a listing of promotions (that is, conferring of titles of doctor, master,
bachelor) of the Jagiellonian University, Krakow, between 1402 and 1849, as compiled by one
J. Muczkowski in 1849 [1][2].

Of potential interest to us as potential relatives of Georg Baresch (a.k.a. Barschius) are mentions of:

    * one “Mart. Barscius Crac.” an “excellent man”, made Master of Arts and Doctor of Philosophy in
1636 (p. 303 of the book, 568 of the PDF)

    * one “Petrus Barschius Crac.”, an “adolescent, carefully examined in the good Arts and
Philosophy” in 1647 (p. 320 of the book, 585 of the PDF).

The “Crac.” following the names is most likely short for “Cracoviensis”, “of Cracow”, thus making it
less likely that these are relatives of our Baresch. Also Petrus Baresch is perhaps too young to be a
brother, if anything he would be a son or nephew.

Dealing with the Latinised names of course brings with it the possibility that several different
vernacular names may map to one Latin name.


[2] the book is mis-titled “Zabytki z dziejów, oświaty i sztuk pięknych, Volume 1”, which is merely
a frontispiece indicating the book is part of a collection of incunabula relating to the arts and
education, in this case held by the Polska Akademja Umiejętności (also incorrectly listed by Google
as the author, who was Muczkowski), a learned society founded in the 19th century. whose name is
rendered in English as “Polish Academy of Learning”, or occasionally as “Polish Academy of Arts and
Sciences” or “Polish Academy of Sciences and Letters”.

From: Jan Hurych
To: Journal of Voynich Studies
Sent: 01-27-2010 10:59:26 AM


Refer also to: J.VS communication #328 (Vol. IV, 19 JAN 2010) " J.VS: The phonetisation of the
Voynich Manuscript text: Voynich text to synthetic speech "; and the subsequent vms-list
thread (launched 20 JAN 2010) " VMs: Phoneticizing Voynich text with synthetic speech ".

The choice of a language and conversion of VM transcript is critical for the phonetisizing. Beside  Latin
and Czech, the English is the next contender. It seems that our conversion of VM transcript to Latin
does not really crack the secret of the VM. It is actually only decoding the arbitrary font of EVA into
Latin text per my Historical Latin Letter Frequency Table, and there is probably another encoding still
present in the VM on top of it. I would like to continue searching for medieval frequency tables -
as I was told there were several of them made at medieval time. In reality, EVA actually provided only
monoalphabetical enciphering of VM letter symbols. By using the table, we simply "deciphered" EVA
as well as any other monoalphabetic substitution cipher if it was present in the VM already. All that of course,
provided the VM plaintext is in Latin. Not so surprisingly - by doing that, we did not get the readable,
sensical plaintext in Latin. That's why I got the idea for phonetisizing the result.

As for other languages, I tried Czech language conversion table but its phonetisizing sounds, just superficially,
of course, similar to the Latin one, most likely because of the sentence structure present in the VM. We may try
the other languages later but for the meantime, the best resources and results we now have are still for Latin,
so I would suggest to stick to that one for the time being.

Pronounciation of medieval Latin is quite similar to International phonetic table while English pronounciation
of the same text is quite different - so what we need is the true phonetic method. Example: letter "a" can be
pronounced in English several ways, depending on the syllable, surrounding consonants and the position of "a"
in the syllable. In Latin, there is only one way, that is as in English "ah".
The English pronounciation of medieval Latin stemmed from the fact they originally vocalized the written word
and did not bother about true Latin pronounciation. So sometimes it completely distorted the sound of the word,
and I heard even priests often pronounce it English way (say suffix "-atio" in Latin is pronounced as "ah-tsio" while
English pronounciation sounds like "ey-shio"). In this case, the phonetic version could lead to erroneous conclusions.
Similar problem would be of course with other languages, typically with French (surprisingly, not so much with
French Latin ;-).
The phonetisizing (by the way, I wanted to use the word "voicing" or "vocalizing" or even "reading" but  we are
doing here something else again) serves several purposes, namely as another tool in the process of decoding the VM.
It is not accurate of course but may be quite inspirational. It is simply just another tool how to crack the VM.
It may help in the search for the original language, the sentence and word structure and/or information flow.
So far the VM researchers worked with written version only, so it is for us completely new game. Also, we cannot
expect any deeply revealing results soon.
The existing text-readers, when they encounter unpronouncable word (unpronouncable in English that is) are just
spelling it letter by letter, again in English spelling of course. It is caused mostly by low frequency of vowels since EVA
provided just the arbitrarily chosen letters. To eliminate that inconvenience, we have to break those words in syllables -
if it helps - or insert some vowels (the best is so called  quiet "e"). We do not know exactly which are the true vowels
in the VM plaintext but I compared it with Sukhotin's results and his vowels basically agree with those in Latin conversion
(not in that order of course, but make it pronouncable all the same). Also, the conversion shows those vowels
(that is a, e, i, o, u, y)  have higher frequency than the EVA arbitrary "vowels", so we are apparently on the right track.
Therefore, for the meantime, we should be able to carry on with the existing conversion until we find the more accurate one.
Still, there are words that remain unpronouncable even in Latin pronounciation, mostly the clusters of syllables. There are two
possible reasons: first is the mentioned inaccuracy of the table of conversion (given the fact the some letter frequencies are
too close to the neighbours to establish the proper order), second is there may be further encoding present in the VM -
say  abbreviations, anagramming, even partial transposition cipher, e.t.c.

So far we found only very few complete words (like several words "saint" which of course should be "sanctus" in Latin,
so I believe it was only coincidence), and also some fractions of Latin words (which may suggest the plaintext was in Latin
but somehow further modified).  
Jan Hurych

From: Berj N. Ensanian KI3U
To: Journal of Voynich Studies
Sent: 02-03-2010 7:58:18 PM

J.VS: The Voynich text speaks: latest experiment converting VMS-text to synthetic speech

Dear Colleagues

As mentioned I have been experimenting upon our colleague Jan Hurych's pioneering foundations of
converting Voynich text to speech by means of speech-synthesis software and associated tools. [1]

I have sent to our J.VS Librarian Greg Stachowski for Library deposit # 28-1-2010-02-02
the recording, 2feb2010_KI3U-VMS-Speech-Expmt.mp3, in mp3 audio-file format, of my latest (yesterday)
experiment [2]. This recording contains the conversion-to-speech of a Voynich text-block of approximately
four dozen consecutive VMS text groups / words.

This kind of work remains tedious, with many experiments necessary to obtain an incremental
improvement. This particular experiment proceeded as follows:

1.) Make a tentative conversion table, mapping Voynich text-glyphs to input-strings suitable for
various speech-synthesis programs.

2.) Try the table on various blocks of Voynich text, and change the mappings, the chosen text-block,
and speech-synthesizer, with the aim of hearing the best obtainable smooth flow of sound resembling
speech which might be a known or unknown language. In other words, go for what sounds closest to
someone actually speaking a real language. This is different from, and less difficult compared with
Jan's initial experiments of keeping the mappings conforming to results which tend specifically
toward Latin-sounding.

3.) If necessary, fine-tweak the resulting recording, but uniformly, with straightforward audio-
processing tools.

In the present experiment I used the SAPI5 TTSAPP speech-synthesizer suggested by our colleague
Dennis Fedak / N3ZCK; the socalled "Microsoft Anna" English voice was used. I then used the Audacity
audio software to do click-removal, so as to obtain a somewhat smoother speech flow.

On hearing the results during the experiment I wondered if this is what spoken language sounded like
in ancient Assyria, Babylon, or even Sumer. I wondered if Gilgamesh lived in a speech-environment
resembling the one suggested in the present mp3 recording. This is an indication of the very
interesting subjective reactions which attend this kind of experimentation. As Jan pointed out it is
at the very least inspirational, and indeed one gets many novel ideas about the VMS.

I noticed, or at least thought I noticed, that Currier's observation that the VMS text-line is a
functional entity, comes across at least as often as not, in my subjective reaction to the flow of
the synthetic speech.

In the present recording there are some utterances which do sound to me like tending toward familiar
words, for example: Christus, datum, textus. At this stage I would give Greek equal consideration to
Latin in the more difficult experimentation of keeping a mapping table conforming to a specific language.

It would seem that the next step up from this level of experimentation is to convert the text into singing.
That would be very exciting to hear, especially if one could find two VMS text-blocks which seemed
related, in the sense of one block holding the lyrics to be sung, while the other holds the accompanying
music in musical notation as per our original VMS-text-to-music experiments. [3]

Berj / KI3U

[1] Refer also to:

[1-a] J.VS communication # 331 (Vol. IV, 27 JAN 2010) :

[1-b] J.VS communication #328 (Vol. IV, 19 JAN 2010)  :
J.VS: The phonetisation of the Voynich Manuscript text: Voynich text to synthetic speech, by Berj / KI3U.

[1-c] vms-list thread (launched 20 JAN 2010 by Berj / KI3U) :
VMs: Phoneticizing Voynich text with synthetic speech


[3] see J.VS comms. # 166 - 172 (Vol. II, February, 2008) :

From: Berj N. Ensanian KI3U
To: Journal of Voynich Studies
Sent: 02-04-2010 11:05:39 PM EST

J.VS: The ancient language Bo dies out

Dear Colleagues

A BBC online news article dated 4 FEB 2010 by Alastair Lawson:


" Last speaker of ancient language of Bo dies in India "

" The last speaker of an ancient language in India's Andaman Islands has died at the age of about
85, a leading linguist has told the BBC. "

" Professor Abbi - who runs the Vanishing Voices of the Great Andamanese (Voga) website - explained:
"After the death of her parents, Boa was the last Bo speaker for 30 to 40 years. "

The article also touches on other Adamanese languages. Quite interesting is a provided audio
recording of the Bo language spoken by the lady who was its last speaker and just died, Boa Sr.
Compare it with our recent synthetic Voynich speech experiments: very different.

Berj / KI3U

From: Berj N. Ensanian KI3U
To: Journal of Voynich Studies
Sent: 02-13-2010 9:58:18 AM EST

J.VS: Hurych on Voynich group frequencies: IS THE VM ENCODED? ( and how to solve it in that case)

Dear Colleagues

Our colleague Jan Hurych has deposited in the J.VS Library his 11 FEB 2010 paper:

IS THE VM ENCODED? ( and how to solve it in that case)

Jan discusses the Voynich text and text-attack methods; here's an excerpt:

" If indeed the VM is using codewords, it also suggests the possible method of solution: instead of
letter frequency, we will use the word frequency in our cracking method. By frequency of course I do
not mean the frequency of the words with the same length, that would be pointless exercise. Instead,
we will count the words that are exactly the same, that is they do have the same order of the same
letters. The "top" words (with highest frequency) then could be replaced by the "top" words of the
language in question. "

Berj / KI3U

From: Berj N. Ensanian KI3U
To: Journal of Voynich Studies
Sent: 02-15-2010 10:57:42 AM EST

J.VS: The history of the Libreria Franceschini

Dear Colleagues

As we know, in 1908 Wilfrid Voynich acquired control of a huge collection of old books, many of them
possibly Risorgimento-confiscated Jesuit volumes, by buying the Libreria Franceschini, Palazzo Borghese,
Via Ghibellina 110, in Florence. [1]

And this acquisition, about three years before the earliest claim (1911) for the discovery of the
Voynich manuscript, seemed to mark Wilfrid's meteoric rise in the antiquarian book field.

Since there is some uncertainty about just where Voynich found the Voynich Manuscript, an Austrian
Castle versus the Villa Mondragone, or perhaps even somewhere else, we have of course been curious
about the inventory of the Libreria Franceschini, and its history.

I've just found what appears to be an article on the origin of the Libreria Franceschini in the
Italian Journal Minerva:


Le origini della libreria Franceschini

Minerva, 11 OCT 1908, Vol. XXVIII N.43, p. 1025.

I can't read the Italian, but the article mentions 50,000 volumes, Voynich, and many other names
which may be investigated. One Alice Zimmern, a writer, is mentioned. There may be more about
Voynich in other issues of Minerva.

Berj / KI3U

[1] see the discussion about this between Xavier Ceccaldi and Dana Scott in the May, 2003 vms-list archives:
VMs: Re: Re: Re: Re: Herbert Garland -- Palazzo Borghese

From: Berj N. Ensanian KI3U
To: Journal of Voynich Studies
Sent: 02-21-2010 11:53:12 PM EST

J.VS: Some Wilfrid Voynich matters across ten years before and after the Libreria Franceshini acquisition

Dear Colleagues

Here's a brief listing of some of the additional information gathered since J.VS comm. #335.
My thanks to our colleague Greg Stachowski for straightening me out on my confusing Helen Zimmern
with her sister Alice, also a writer, thus enabling finding the 1908 Pall Mall article referenced in the
1908 Minerva article (J.VS comm. #335, J.VS: The history of the Libreria Franceschini, by Berj / KI3U,
Vol. IV, 15 FEB 2010).

As our colleague Jan Hurych pointed out, Wilfrid Voynich made the Libreria Franceschini his
Italian-location headquarters: this is immediately seen on the cover of Wilfrid's Catalog No. 31,
"An Illustrated Catalogue of Remarkable Incunabula, many with Woodcuts and a Specimen of an Unknown
Xylographical Press, offered by Wilfrid M. Voynich", bearing also the cat-and-mouse emblem, and listing,
besides his London main headquarters, among his Branch Houses (Paris, Florence, Warsaw, Vienna) the
Florence address as: Palazo Borghese, Via Ghibellina, 110 [1].

Greg found "Stead's 'The Review of Reviews' for Australasia", 1908, wherein:

In another article Miss Helen Zimmern describes an unworked mine in the shape of a bibliomaniac's
hoard at Florence. This bibliomaniac bought indiscriminately everything brought to his shop.
Quantity, not quality, attracted him. He bought and bought and stored and stored with a view to the
future, being of opinion that this was the best way to furnish his children with a patrimony which
would repay them a hundredfold. His name was Pietro Franceschini. Among his friends was Carducci the
poet, and Gladstone was a frequent visitor when in Florence. When Franceschini died, his son, the heir of
the congested bookstore, an engineer, who cared nothing for books, sold the concern to the bibliographer,
Mr. Voynich. The agglomeration of years was investigated. Many finds came to light, especially valuable
first editions of English works, for there seems no place like Italy for picking up English curiosities ; and other
treasures are being disinterred daily by the indefatigable miners. "

Jan and our colleague Dana Scott recalled G. Orioli in his 1938 "Adventures Of A Bookseller"
commenting on Voynich and his Libreria Franceschini - that excerpt is readily available in Dana's 19 JUL 2004
vms-list post:

Orioli's characterization of Wilfrid Voynich, delivered eight years after Voynich's death, seems quite at odds with
the very positive characterizations referenced in J.VS comm. #309 (Vol. III, 8 NOV 2009)

Now, the English-language article by Helen Zimmern referenced both in the Italian-language Minerva article and
the above Review of Reviews, is:


This article is within Pall Mall Magazine's Volume 42, and is available for download via google books.

Helen Zimmern (1846-1934) was a highly multi-lingual writer, and early on in her career she
collaborated with her sister Alice on works much appreciated by scholars of the time [2].
Mrs. Aubrey Le Blond, i.e. Elizabeth Hawkins-Whitshed (1860-1934), was a famous mountaineer, author,
photographer, and apparently the world's first lady cinematographer / film-maker. So it appears that these
two interesting ladies crossed into Wilfrid Voynich's cosmic-grade circle at least once.

Helen's article is eight full pages long and includes half a dozen photographs by Aubrey, the first
and last of which are full-page shots of inside the cavernous Libreria Franceschini - the last being
an intriguing shot of one of the socalled "dark rooms". Helen tells the story of Pietro Franceschini [3] and
how, starting with a barrow, he amassed his vast collection of books for the intended cultural benefit of his
children, the collection over the years growing to phenomenal size, and taking as its permanent home the
form of a bookshop near Florence's Bargello on the Via Ghibellina.

This shop was somewhat obscure, but captured the attention of many people of consequence, indeed one
story having it that even Queen Victoria visited incognito. Pietro's cultured daughter's health suffered and so
she could not continue with the Library; her brother, an engineer, was not interested, and after their father died,
he sold it to Wilfrid Voynich.

Quick key-notes on Helen's article:

1-1.) Helen explains why so many early and rare English works are findable in Florence.
1-2. She refers to Voynich as " well-known bibliographer and authority on early books " - this is in
1908 already, and many other references to Voynich from back then confirm it to be quite true.
1-3.) After buying the library from Franceschini's son, Voynich with his assistants began months of
going through the holdings, still unfinished at the time of Helen's article.
1-4.) " Books, pamphlets, MSS. in every known and unknown tongue are being disinterred daily by the
indefatigable miners. " - note "unknown tongue"
1-5.) Antwerp 1554 or earlier Bellero cat-and-mouse map - note Voynich's cat-and-mouse coat of arms,
apparently orginally of the Sessa family of Venice 14th-15th c.
1-6.) " Babel of languages and books "
1-7.) " a book that had belonged to King Charles II. "
1-8.) " Two old Floretine playing-cards "
1-9.) " pamphlets ..... very early scientific treatises "
1-10.) " fine Dantes and Petrarchs "
1-11.) " rare woodcuts of the fifteenth and sixteenth centuries, books whose bindings tell that they
belonged to kings and popes; "
1-12.) " .... but ornamented on its cover by no less an artist than the great Goya. "
1-13.) " in great proportion the books are uncut and in their original bindings " - note it is conceivable that
Wilfrid rebound the VMS with a binding salvaged from the Franceschini hoard.
1-14.) " the quaintest and most valuable discoveries in the backs of the old bindings. "
1-15.) " Voynich never discards a book till he has stripped it of its cover. "
1-16.) " And to think all this is but a scratching of the surface! "

Jan noted that Helen Zimmern wrote that Pietro had died eighteen months before, and so he suggested
we try to bracket the time during which Wilfrid Voynich bought the Libreria Franceschini. In reply I
sketched this scenario:

2-1.) Assume that this issue ( OCT 1908 ) Pall Mall appeared in SEP 1908.
2-2.) Assume there was a 3-months lead time to get an article into any particular issue;
hence Zimmern's article, intended for the OCT 1908 isssue, reached Pall Mall's editorial offices in JUN 1908.
2-3.) Assume that the "eighteen months before" can be referenced to JUN 1908.
2-4.) Therefore the month of death of Pietro Franceschini would be around DEC 1906 - JAN 1907.
2-5.) After Pietro died, Helen says his son sold some items merely by paper weight, before he encountered
Wilfrid and sold it all.

So, it looks like 1907-1908 is when Wilfrid acquired the Libreria Franceschini, certainly well before the earliest
claim (1911) for the discovery of the VMS.

Jan also noted items 1-14.) and 1-15.) in the key-notes above and commented again on the old oddity of Wilfrid
claiming that when he first found the Voynich Manuscript he initially missed Marci's letter. [4]

The earliest direct indication of Wilfrid Voynich engaged in the book trade which I have found is an
advertisement appearing in a series of issues of the newspaper "THE ACADEMY", beginning with the
Saturday, July 30, 1898, issue. A picture of this advertisement is available in the J.VS Library in
deposit # 29-1-2010-02-21 :

In this advertisement we see Voynich offering:


We can't help notice how similar this 1898 list is to some of the classes of material in the Franceshini described
by Helen Zimmern ten years later: incunabili, early Americana, copies with expurgated passages doubtless judged
obscene or sacriligeous, old English books, good stuff and rubbish, monastic account-books. Therefore, it is
natural to wonder if Wilfrid Voynich had been one of Pietro Franceschini's customers as far back as 1898,
obtaining from Pietro priceless treasures at Pietro's characteristically generous prices, while also making
connections with people of consequence among Pietro's circle. And perhaps Voynich waited, and when the
opportunity came he became the heir of Pietro's fabulous hoard.

Where-ever, and whenever, Voynich obtained the Voynich Manuscript and associated matter, it is
reasonable to suppose that the Libreria Franceschini might have been involved in some manner or
other. We know that during his research on his mysterious cipher manuscript Wilfrid Voynich
contacted Bohemian archivists for help - can we establish when this occurred at the earliest?
Was it perhaps before 1911?

Let us frame around Voynich's acquisition of the Libreria Franceschini his early publications, exhibits, and lectures.
The list of these I've compiled from searching online follows in Table 336-1. I do not know what might be missing
from this list, and which might be duplicates.

--------------- TABLE 336-1

Voynich's dated output is sorted in order. Un-sorted is listed last.

Names appearing with Voynich, apparently his book-trade associates, are:

*) 1898 : Wilfrid M. Voynich & C.A. Edgell (Firm), Wilfrid M. Voynich, C. A. Edgell - Antiquarian booksellers;
{ apparently Edgell received an M.A. from Cambridge in 1898, and did some singing there }

*) 1902 : Voynich, Wilfrid M., firm, booksellers, London, Ernest George Ravenstein - Maps;
{ Ravenstein appears to have been a long-established authority on geography, in particular Russian }

*) 1903-1904 : Wilfrid M. Voynich, London, George Mackey, Birmingham


*) A first list of books
Wilfrid M. Voynich & C.A. Edgell (Firm), Wilfrid M. Voynich, C. A. Edgell - Antiquarian booksellers - 1898 - 66 pages

*) A first [-sixth] list of books offered for sale
Wilfrid M. Voynich, firm - 1900

*) A second list of books
Wilfrid M. Voynich (Firm) - Antiquarian booksellers - 1900 - 212 pages

*) List of books offered for sale at the net prices affixed
Wilfrid M. Voynich (firm) - Rare books - 1900

*) List of books offered for sale at the net prices affixed
Voynich, Wilfrid M., firm - 1900

*) A third list of books
Wilfrid M. Voynich (Firm) - Antiquarian booksellers - 1901 - 160 pages

*) A fourth list of books
Wilfrid M. Voynich, London - 1901

*) A fifth list of books
Wilfrid M. Voynich (Firm) - Antiquarian booksellers - 1901 - 86 pages

*) A sixth list of books
Wilfrid M. Voynich (Firm) - Antiquarian booksellers - 1901 - 739 pages

*) A first list of books (second edition).
Wilfrid M. Voynich, London - 1902

*) A seventh list of books
Wilfrid M. Voynich, London - 1902

*) An eighth list of books on exhibition by W.M. Voynich: I Soho Square, London
Wilfrid M. Voynich (Firm) - Antiquarian booksellers - 1902 - 897 pages

*) Supplement to the eighth list of books ...: consists of an unknown map, ....
Voynich, Wilfrid M., firm, booksellers, London, Ernest George Ravenstein - Maps - 1902 - 958 pages

*) A ninth list of books
Wilfrid M. Voynich (Firm) - Antiquarian booksellers - 1902 - 1117 pages

*) Short catalogue of Second-hand books offered at the net prices affixed by ...
Wilfrid M. Voynich - 1903

*) Second-hand books and manuscripts
Wilfrid M. Voynich, London, George Mackey, Birmingham - 1903

*) Second-hand books
Wilfrid M. Voynich, London, George Mackey, Birmingham - 1904

*) Wilfrid M. Voynich catalogue
Wilfrid M. Voynich - 1910

*) Early works on pure and applied science - part I - A-L , with index of subjects
Wilfrid M. Voynich, London - 1912

*) Remarkable incunabula, many with woodcuts, and a specimen of an unknown ...
Wilfrid M. Voynich, London - 1913. { This may be Catalogue No. 31 - see [1] }

*) Catalogue
Wilfrid M. Voynich - 1914

*) Exhibition of manuscripts and early printed books from the collection of W.M. Voynich;
Art Institute of Chicago, October 7 to November 3, 1915. { This is the first public exhibition of the
"Roger Bacon Cipher Manuscript" which nowadays we call the Voynich Manuscript. [5] }

*) Lecture: How to study fifteenth century books, by Wilfrid Voynich at the Art Institute of Chicago,
January 16, 1917. [5]. { This is interesting in light of the recent radio-carbon dating of some pieces of
VMS parchment to 1404-1438. }

Un-sorted Wilfrid Voynich output:

*) Valturius: De re militari : a XVth century manuscript executed at Rimini in ...
Wilfrid M. Voynich - Illumination of books and manuscripts, Italian - 10 pages

*) Early works on English history, literature and science - part 2 - E-L.
Wilfrid M. Voynich, London

*) Early printed books, many with woodcuts, and in fine bindings
Wilfrid M. Voynich, London

*) Early printed books, many with woodcuts -
Wilfrid M. Voynich, London

*) Rare books, printed in the XV, XVI and XVII centuries, not to be found in ...
Wilfrid M. Voynich, London

*) A catalogue of rare books, printed in the XV., XVI. and XVII. Centuries, not ...
Wilfrid M. Voynich (Firm) - Antiques & Collectibles - 19?? - 116 pages

*) An illustrated catalogue [of rare and early printed books]
Wilfrid M. Voynich (London) - Rare books

*) An illustrated catalogue of remarkable incunabula, many with woodcuts, and a ...
Voynich, Wilfrid M., Firm - Bibliography - 19?? - 178 pages

--------------- End TABLE 336-1

We note from Table 336-1 that in 1912 Voynich published on early works on pure and applied science.
This is the same period during which the VMS was allegedly found, and we recall that Voynich
believed the VMS to be a scientific encyclopedia by Roger Bacon. We also recall learning from Helen
Zimmern's 1908 article that the Libreria Franceschini was loaded with rare early scientific works,
often in pamphlet form (see 1-9. above).

The main objective in this communication has been to organize early Wilfrid Voynich and VMS data
within the perspective of his acquisition of the Libreria Franceschini, so as to facilitate gauging the role,
if any, played by the Franceshini in the Voynich Manuscript story. That this notion is quite reasonable
was evident during the recent J.VS teleconference of 19 FEB 2010 (participants: Jan Hurych,
Richard SantaColoma, Berj / KI3U) where among the altogether dozen VMS topics we discussed:

3-1.) The likelihood of blank old parchments to be found in Franceshini's "dark rooms".
3-2.) Wilfrid possibly obtaining manuscripts from Pietro Franceschini before Pietro died.
3-3.) Old information that Wilfrid bought the Libreria Franceschini in 1909, is wrong - he acquired it sometime 1907-1908.

Subsequent to the teleconference the off-J discussions flowing from the above branched out into some
discussion of parchments per se: parchment-making, parchment properties, and palimpsests. Here's a
paper on history of parchment making, by Meliora di Curci, which we noted:

The History and Technology of Parchment Making

The alignment of fibres in the collagen which makes up the parchment seems to offer clues about the
parchment's history. One interesting item di Curci writes is this:

" Low Relative Humidity Prolonged exposure to an environment under 40% relative humidity will dry
out parchment and bring its water content to below 10%. Once again this is a slow process which may
take several months or years. Prolonged exposure to this environment will turn the parchment harder
and harsher, eventually cracks would appear in the surface and the inks and paints will detach (Reed, 1975, p. 94). "

The inks and paints will detach she says. Hence perhaps such a parchment could be restored and re-used,
if found otherwise undamaged.

The idea of the VMS being, at least partly a palimpsest, has been pretty much discounted over the years.
Jorge Stolfi, in passing, brought up the idea of the VMS's f68v3 "T-O map" as perhaps a palimpsest back in 2002
in a vms-list reply to Nick Pelling [6].

I think that maybe the idea does deserve some attention - after all the existence of the alleged "Tepenecz" and
other barely legible writing on VMS f1r would seem to invite at least a consideration of that folio having been a re-use.
The main arguement against VMS palimpsest these days seems to be that the VMS folios show no indications of
scraping - a strange idea when we consider that scraping is a major step in the manufacture of parchment.
I haven't scraped parchment, but I've scraped plenty of old really-in-bad-condition suede leather, and I've been astonished,
completely surprised, at how fresh and lustrously new condition the old suede can be restored to with serious dedicated
effort and patience.

Elsewhere on the net, the discussions about palimpsests seem to indicate that there were processes which irreversibly
removed old writing from parchments - i.e. apparently no faint traces would later re-appear. I really do not know enough
about this subject, but it seems worth looking into a bit more, rather than dismissing it summarily.

Finally, the above data surrounding Voynich's acquisition of the Libreria Franceschini, when followed up on in detail,
again confronts us with the almost unbelievably far-and-wide high-quality spectrum of contacts Wilfrid Voynich enjoyed
around the world. One thing which I continue to notice is that some among Voynich's admirers when they speak of him,
seem to suggest a sixth sense of sorts in the man Voynich. And one wonders if Wilfrid and Ethel Voynich, and perhaps also
Miss Nill, were themselves interested in such notions, notions of the paranormal. In their heyday it would not have been
necessarily frowned upon if the Voynich's had participated in seances, say. One could fathom Wilfrid attemtpting to make
contact with the spirit of the VMS's author - presumably in his case Roger Bacon.

Anyway, the closest I've come to finding any evidence of the Voynich's involved with any paranormal investigations is
Dr. Erla Rodakiewicz (nee Erla B. Hittle). We recall that in her July 17, 1945 letter to Dr. Leonell Strong [7], Erla tells him
that Wilfrid was one of her closest friends, and that she had met him in Italy in 1898. Well it turns out that Erla in 1911 was
living in Austria and was a member of the Society for Psychical Research, the original and prestigous SPR - she is so listed in
the SPR's proceedings that year. I think she may have been in the SPR some other years also. Interesting then, that someone
so close to Wilfrid was in the SPR.

Dana added to the Erla Rodakiewicz data this interesting item:

The Voynich MS
Letter of Friedrich von Hayek to Eric Sams

where we learn of a cocktail party in 1923 given by Erla, including among her guests the anthropologist Franz Boas.
Wilfrid was there, and he apparently astonished the young Friedrich von Haek with his phenomenal memory and knowledge.

Berj / KI3U

[1] Rafal T. Prinke dates the Voynich No. 31 Catalogue in the period 1903 - 1915 : vms-list post "Re: VMs: Voynich Catalogue No. 31", 26 JAN 2003
See Table 336-1 for a possible 1913 dating of this Catalogue.

[2] The sisters Zimmern cooperated at least in the early 1880's on literary pursuits - they co-produced
"Half-Hours with Foreign Novelists". Perhaps they eventually got to know Ethel Voynich.

[3] We must not confuse our Pietro Franceschini with another 19th c. Pietro Franceschini, who
apparently was a scholar of some sort.

[4] The uncertainty about Marci's 1665/1666 letter, allegedly accompanying the Voynich Manuscript
when found, is reflected in D'Imperio, pages 1-3, where she gives the two versions, "found between
the pages of the manuscript", and "attached to the front cover". Marci's letter was clearly enveloped in
the usual manner and wax-sealed: Jan comments that presumably, according to standard VMS-history
indications, it would then have been Athanasius Kircher who broke the seal and opened the letter since it
was written to him, but why then was this letter not Jesuit property?

On those same pages D'Imperio reflects another uncertainty in the VMS story, namely where Voynich
found his manuscript: at the Villa Mondragone, or a castle.

[5] The Bulletin of the Art Institute of Chicago, Volumes 1-12, is available now online via google
books (~ 40.8 MB). It contains a half a dozen or so notices of Wilfrid Voynich's activities.

[6] vms-list post "VMs: Re: f68v3 ("Andromeda") and it's T-O map...?", Jorge Stolfi, 28 SEP 2002.

[7] Glen Claston (GC) provides online the VMS correspondence of Dr. Strong at Voynich Central here:

From: Berj N. Ensanian KI3U
To: Journal of Voynich Studies
Sent: 02-25-2010 11:57:09 PM EST

J.VS: Voynich and connections with Psychical Research

Dear Colleagues

In J.VS communication #336 [1] I brought up the investigating of any connections between Wilfrid and
Ethel Voynich, as well as Miss Anne M. Nill, with psychical research. This is not to be confused
with long-plowed investigations of possible Dee and Kelley-as-medium theories of the origin of the
Voynich Manuscript.

Why look into possible psychical activities on the part of the major players in the early story of
the Voynich Manuscript?

The most critical answer is that if indeed any of the early major players were more than casually
involved in paranormal work, then perhaps they experimentally employed such work in their approach
to understanding the mystery of the VMS, and if we can find any records of such experiments, we may
be able to glean some clue or other of what these major players may have known or thought about the
VMS but never publicized.

So for example, if Wilfrid and Ethel Voynich ever sat in a seance at which supposedly some
intelligent and willing spirit-librarian "came through", and Wilfrid asked that spirit some
questions about his Roger Bacon Cipher Manuscript, then surely we would be very interested in just
what questions Wilfrid asked. Would we not?

It is clearly beside the point here what we personally, or what religion and science, think about
seances and such, rather the essential point is that this area of investigation may yield new
information about what those early players most in the know about the VMS, were thinking and
pursuing about it. Needless to say we would likewise be interested in any mediumistic-like notes of
Baresch, Marci, Kircher, etc., at least as much as in anything Kelley-the-medium.

Some folios from the VMS went missing during the 20th century - can we find out something about
their fate along this present track? Were they perhaps used in a seance, perhaps a seance which
required a "sacrifice" - in this case the cutting out of a folio from the VMS? That's not such a
far-fetched thought when we consider that some of the VMS illustrations, like f72v2 and f82r, can
easily give the impression of seance, trance-induction, trance medium-ship, and astral-projection
and related techniques, and further that the nine-rosettes foldout can be imagined as a complex
mandala-network. It is not impossible that a folio in the VMS indicated, in some understandable
manner or other, that in order to lift the veil of the book's mystery, this folio should be cut out
and sacrificed, perhaps by burning, in a seance.

In #336 we saw that the well-connected Dr. Erla Rodakiewicz, who had met Wilfrid Voynich in Italy in
1898 and counted him as one of her closest friends, was a member of the original (1882) and
prestigous Society for Psychical Research (SPR) at least in the year 1911, when she was living in Austria.

Dr. William Romaine Newbold, most notably of the University of Pennsylvania, Wilfrid Voynich's
closest collaborator in the effort to demonstrate the Roger Bacon Cipher Manuscript (Voynich Manuscript, VMS)
as the work of Roger Bacon, was not only a long-time associate of the SPR, in particular the American
incarnation (ASPR), but prominent enough early on so as to have had the lead-founder of the SPR,
Frederic William Henry Myers, read a Newbold paper at an SPR meeting early in 1896 [2].

Newbold's extensive involvement in the world of paranormal research is curiously absent from his
confusedly-written biography still currently presented online by the University of Pennsylvania,
which does however remark at length on Newbold's VMS work [3]. In contrast, the 1917 General Alumni
Catalogue of the University of Pennsylvania is not so shy in its brief sketch of Newbold.

Newbold appears with some regularity in the pages of psychical / paranormal literature from at least
1895, that is the year before he became Dean of the Graduate School at U. Penn, until his death in
1926. He participated in experimental seances and tested the genuineness of spirits raised in seances.
We can learn quite a bit about Newbold's attitudes toward personal and scientific evidence
from reading some of his psychical commentary.

For example, in the Proceedings of the Society for Psychical Research, Part XXXIII, Supplement, Reviews,
Newbold at length reviews (dated November 11th, 1896) Nevius's 1896 "Demon Possession and Allied Themes",
ending his comments with:

" Dr. Nevius failed to see that the evidence which is sufficient to bring a mass of phenomena under
an accepted category may be wholly insufficient to establish the category. The evidence which he has
collected may be subsumed under several categories with equal ease, but there is nothing in the
evidence to determine us to the choice of the spiritistic rather than of another. That must be done
by the more complex considerations that fix for each man his attitude towards the Universe. "

Also in Part XXXIII, Supplement, under "Psychical Research in American Universities. By Harlow Gale",
Newbold is quoted:

" My own interest in the matter is very keen, and I regret very much that my work leaves me no time
for original research. I do not, however, introduce it into my teaching. In the first place I am
giving no courses in which the subject would naturally find its place, and in the second, I am by no
means sure that it is desirable, either for the sake of the cause or for the sake of the students,
to do so. I always, however, so far as possible, endeavour to arouse the interest of any individual
student who seems possessed of the good sense and judgment which are so essential in one who intends
to deal with such questions. "

One technique which Newbold (and other psychical researchers) employed during a seance to test a
"spirit" claiming a certain identity, was to give it an expression in some language which the
trance-medium facilitating the contact with the spirit did not know, say Greek if that spirit should
have understood Greek. These experiments sometimes produced responses from the spirit with "odd"
executions of the languages in question.

Now then, it seems entirely conceivable that Newbold would have gotten the idea sometime during such
a seance experiment, where there was contact with a spirit claiming to be able to bring back lost
knowledge, to put in front of the spirit (in actual practice the trance-medium) a sample of the strange
VMS script, and ask: where is this from, who wrote like this, and when and where?

And now the crucial point: assuming the spirit came back with some answer which was not immediately
useless or ridiculous, Newbold's NEXT follow-up question((s) would likely reflect what he, and perhaps
Wilfrid Voynich too, thought about the VMS at the time, and such thoughts of Newbold and Voynich
may never have been publicized. But surely we would be interested in what those thoughts were!

It was standard practice in experimental seances to make a careful record of what was transpiring.
Hence Newbold's surviving private papers specific to psychical research would seem to be worth
combing for any hint of something like the above involving the VMS. If he did put some VMS script
before a spirit during a seance, which particular VMS text was it? Why that text in particular?

In general the surviving private papers specific to psychical research of anyone close to the Voynich's
in the early VMS days, would be worth a combing through to see if anything VMS turns up.
So far as I know, this is virgin VMS territory.

Of course this means getting a better detailed picture of Wilfrid Voynich's tremendous world-wide
web of contacts, an ongoing job important for Voynich work in general anyway.

Judging from his July 1921 article in Harper's on the Roger Bacon Most Mysterious Manuscript,
John M. Manly seems to have been familiar with Newbold's Hilprecht-subconscious-reasoning paper [2].
Manly so far doesn't show up on the psychical radar, and he doesn't really seem the type, but perhaps he
kept tabs on the field via his brother Charles M. Manly who worked with Samuel Pierpont Langley on the
Langley Aerodrome - Langley was a Vice-President of the SPR at the time of his death in 1906.

Turning to Ethel Lilian Voynich, the daughter of George Boole and Mary Everest Boole,
Rafal T. Prinke has available online an excellent pdf document compiled by him in 2002:

Genealogia Booleana

Therein Prinke writes in the entry for Ethel Voynich's mother :

" Mary Everest [1832-1916] pedagogist, mathematician (invented string geometry) and spiritist (wrote
The Message of Psychic Science for Mothers and Nurses, besides many other books) "

Indeed, ELV's mother's extensive output on the paranormal is easily found by searching the internet.
In October, 1908, the same month Helen Zimmern's article in Pall Mall on the Libreria Franceschini
appeared [1], Mary Everest Boole wrote in the Preface of her book, "The message of psychic science
to the world", as follows:

" My father, T. R. Everest, was a learned occultist in days when occultists were few. He did
everything in his power to call attention to the dangerous re-action which must come if the clerical
and medical professions persisted in ignoring the phenomena of Mesmerism, Trance and Clairvoyance. "

" Preface to Private Edition. In 1883 a small edition of the book was published, but not advertised.
Very few copies of it were ever sold. The preface to that edition was as follows :— .............
A society has lately been formed, called the Society for Psychical Research, which announces itself as
willing to receive and examine, and as far as possible to classify, evidence on such subjects as
thought-reading, clairvoyance, apparitions, and haunted houses. It is not difficult to foresee that
the future history of this new movement will in many respects repeat the experience of the past.
The leaders, many of whom are men of unquestioned ability, will pursue their course in a spirit of calm
and patient inquiry, equally unmoved by the. satire and antagonism, and by the over-excited
curiosity and too ready belief, which will seethe around them. They will combine and utilise the
work of isolated observers ; and in due time the world at large will profit greatly by their
labours. But meanwhile a terrible amount of quite needless suffering will be caused to those who
take no part in the movement, by the mere unsettling of their ordinary habits of thought. And the
occasion may perhaps not be inopportune to remind young women who have not the leisure or the
aptitude for systematic investigation, of the same principles which were so much forgotten by
unscientific readers in the first shock of surprise caused by the publication of the " Origin of Species." "

With Erla Rodakiewicz, William Romaine Newbold, and Ethel Lilian Voynich's own mother openly
involved in the paranormal field, we have Wilfrid and Ethel Voynich well "surrounded" by at least
the opportunities to experiment paranormally with the VMS mystery. So far as I can presently
ascertain, it is generally unknown if they did so. It's a point which should be resolved, and if it
turns out that they did psychically experiment involving the Voynich Manuscript, then it behooves us
to find out the details - we may well learn something quite interesting in the hidden VMS story.

Berj / KI3U

[1] J.VS comm. #336 (Vol. IV, 21 FEB 2010):
J.VS: Some Wilfrid Voynich matters across ten years before and after the Libreria Franceshini acquisition, by Berj / KI3U.

[2] On 28 DEC 1895 Newbold presented at a meeting of The American Psychological Association:
A Case of Dream Reasoning. This is reported in The American Naturalist, Volume 30, February 1896, p. 170;
The PSYCHOLOGY department of this magazine was edited by Newbold. Newbold's lecture was apparently
attended by William James. Two of the three cases Newbold delt with were Hilprecht's famous Assyriology dreams
(i.e. famous in Assyriology; Hermann Volrath Hilprecht was one of the major figures at the dawn of U.S. Oriental-Archeology
which began with the Babylonian expeditions organized by the University of Pennsylvania in the late 1880's.
At U. Penn Hilprecht and Newbold were friends.).

In the quarterly review and index BORDERLAND, Vol. III, No. 2, April, 1896, on page 224, is reported
that at the last meeting of the Society for Psychical Research, Mr. Myers read an exceedingly interesting paper
on Sub-Conscious Reasoning by Newbold (the Hilprecht case) and it is to appear in the forthcoming issue of the
Proceedings of the SPR.


From: Berj N. Ensanian KI3U
To: Journal of Voynich Studies
Sent: 03-23-2010 11:34:11 AM EDT

J.VS: Historical Resources Search Engine

Dear Colleagues

Here's a 23 MAR 2010 dated article by Mark Ward on the BBC online news service:

" Search engine collects historical resources "

" A search engine is being created to help historians find useful sources. The Connected History
project will link up currently separate databases of source materials. Once complete, it will
give academics or members of the public a single site that lets them search all the collections.
Once completed the search engine will index digitised books, newspapers, manuscripts,
genealogical records, maps and images that date from 1500-1900. "

" In general, said Professor Shoemaker, the different collections possess different types of
materials so there is little overlap between them. Currently 12 institutions have signed up to
contribute their collections but more are expected to join in the future. The initial partners
include the University of Sheffield, the Institute of Historical Research, the University of
Hertfordshire and King's College, London. The first phase of the Connected History project
should be completed by March 2011. "

Berj / KI3U

From: Greg Stachowski
To: Journal of Voynich Studies
Sent: 04-07-2010 5:16:21 PM EDT

J.VS: Speculations on dyslexia as a source of oddities in VMS text

Robert Teague's recent efforts attempting to decode the star labels in the VMS, for example those
on f68r3, involve in part insertion of missing letters, deciphering anagrams and alternative
letter substitutions. The latter appear to follow some underlying structure, as first noted by
Philip Neal and expanded on by Robert. These features have come under close scrutiny due to the
potential ambiguity they introduce into a conventional encryption and decryption system.

Without commenting here on the validity of Robert's results, it seems worth noting that such
features might be explicable as the effect of some form of dyslexia affecting the VMS author.
No statistics are available for dyslexia in the likely time period, but judging by the modern level
of incidence the underlying neurological condition can't have been unlikely, being masked instead
by the general high level of illiteracy and lack of an established orthography.

If we assume, for the sake of this discussion, that Robert's decodings are indeed correct then
this might explain the mixed up letters: what looks like anagramming to us might in fact have
looked fine to the author, in the same way as modern dyslexics are often unable to see the
mistakes they make.

Possibly, statistical analysis which takes into account the patterns arising in writing by
dyslexics might be used to test this idea. Perhaps even the raw (not decoded) VMS text may
exhibit such patterns; speculating very optimistically, these patterns might even explain some of
the departures from conventional language statistics. No such study has (as far as I know) been
carried out.


From: Berj N. Ensanian KI3U
To: Journal of Voynich Studies
Sent: 04-10-2010 11:40:18 AM EDT

J.VS: Re: Speculations on dyslexia as a source of oddities in VMS text

[ Redacted from off-J discussions 8 - 9 APR 2010 ]

Robert Teague says:

I've been doing a bit of research. The part of dyslexia dealing with writing is dysgraphia.
It has to do with understanding the directionality of letters and words.
It was hard finding specifics; many of the websites quoted each other. But here is some:
The dyslexic may reverse letters like b and d, or p and q, either when reading or writing.
He may invert letters, reading or writing n as u, m as w, d as q, p as b, f as t.
Another website mentioned dropping the letter A.

Berj / KI3U says:

Over the years I've a few times encountered dysgraphia in radiotelegraph communications.

Robert says:

My impression of the VMs letter set is that it is relentlessly left to right,
(dyslexics have trouble remembering which way to write.) and all the
letters are easily distinguished from each other with no reversed or
inverted letters.
I find the script simple, easy to remember, and elegant.
The word-initial and word-final letters might have helped the Author
with spelling (or at least he thought so).

Greg Stachowski says:

Web sites about dyslexia are sometimes confusing because they try to describe all possible
symptoms. Not all dyslexics exhibit all possible dyslexic traits; for example not all dyslexics
have trouble remembering which way to write. Some just have trouble with spelling, some flip
letters, some appear to read and write fine, but can't distinguish left from right. Which is a
particularly interesting example - it's like the mental association between the words "left" and
"right" and the concept of leftness and rightness is permanently broken. They will write left-to
-right quite happily, drive on the correct side of the road, and so on, but if you say "left"
they have no idea what you mean and despite being intelligent and capable can't learn it. Many
milder dyslexics (i.e. the majority) learn to mask the traits very well, though sometimes through
considerable effort.

Anyway, I would be wary of placing too much weight on this idea as an a priori explanation of the
features in the VMS, in other words it should follow from the data rather than leading the analysis.

Berj says:

Oh, agreed. But nevertheless the idea seems to open up new possibilities for analytic models of
the VMS text, and just in that it may prove fruitful, even if the eventual conclusion is that it
is unlikely that the VMS text writer was innately dyslexic / dysgraphic. I suppose it is also a
remote possibility, very remote, that he/she was intentionally mimicing dysgraphia.

Robert says:

I think the idea is a good one, needs to be kept available for follow up.

Berj says:

Indeed so. From the earlier examples you guys gave of dyslexic symptoms it seems especially
interesting to exploit this idea with the problem of the many almost-but-not-quite-the-same
Voynich words / groups.

[ end J.VS comm. #340 ]

From: Berj N. Ensanian KI3U
To: Journal of Voynich Studies
Sent: 04-21-2010 2:28:26 PM EDT

J.VS: Rector Martinus Santinus

Dear Colleagues

Returning for a moment to the problem of finding out more about Padre Martinus Santinus (see J.VS
communication #251, Vol. III, 11 FEB 2009, etc.), and via him possibly finding connections
between Baresch, Horczicky / Sinapius/ Tepenecz, and the VMS, here's a piece of information:

Geschichte der Prager Universitaet zur Feier der fuenfhundertjaehrigen Gruendung
derselben verfasst von Wenzel Wladiwoj Tomek
Prag, 1849

which is Tomek's History of the Prague University for the celebration of its 500 years founding.

A copy of this book is available online via google books.

On page 362, paragraph f), Tomek lists the Rectors of the Jesuit college at St. Clemens, as well
as Rectors of the annexed Karl-Ferdinand University, over the period 1622-1638, and among the
seven names is given:

1626 - 1629. Martin Santinus.

We've long known Santinus was a Rector of Jesuit Colleges at Prague and Lemberg, but I hadn't
seen the specific dates before, so I thought I'd communicate them here for easy reference.

Berj / KI3U

From: Greg Stachowski
To: Journal of Voynich Studies
Sent: 04-28-2010 6:55:51 PM

J.VS: Vatican Library to digitise manuscripts

Nice to see they are using a sensible image format.

I wonder what nuggets of Kircher, or better the VMS, might come up.


From: Berj N. Ensanian KI3U
To: Journal of Voynich Studies
Sent: 06-1-2010 10:05:00 AM

J.VS: Plato Code

Dear Colleagues

Here's a BBC online news article dayed 29 JUN 2010 :

" Manchester historian deciphers hidden 'Plato Code' "

" A science historian in Manchester claims to have deciphered secret messages hidden in the ancient writings of the philosopher Plato.
Dr Jay Kennedy from the University of Manchester has revealed that the legendary Greek philosopher Plato used a regular pattern
of symbols to give his books a musical structure. Plato's books had a key role in establishing the foundations of Western culture.
But the existence of a so-called 'Plato Code' has long been disputed. "

" Working from the original scripts, he observed that Plato used a regular pattern of symbols, inherited from the ancient followers of
Pythagoras, to give his books a structure hidden beneath the surfrace. "

" The hidden codes reveal that Plato anticipated the Scientific Revolution 2,000 years before Isaac Newton, discovering its most
important idea - the book of nature is written in the language of mathematics.

"In ancient times, many of his followers said his writings were written in symbols; in modern times that was denied," he said. "

Berj / KI3U

From: Berj N. Ensanian KI3U
To: Journal of Voynich Studies
Sent: 07-14-2010 1:06:10 AM

J.VS: Definitions for n-graph spectroscopy

Dear Colleagues

As you know, a couple of months ago starting May 3rd, our off-J discussions focused on some
new approaches to analysis of text. Two related major avenues of work have been following

{1.} Generating for test purposes in text analysis pseudo-random blocks of "text", steerable
as to block gross-characteristics like: average group-size.

{2.} Formalizing the study of "text" n-graphs : n-graph spectroscopy

I've put text in quotes to remind that in general, sequences of groups of symbols, or glyphs or
letters or characters are, in the analytic work, being regarded as "text".

Both of the above were simultaneously hitched to the development of a computer program which
received the name " JVSCGS ", standing for " J.VS Chaotic Groups Generator ", which is designed
to facilitate both the above, and will soon be made available to others in the J.VS Library.
A little bit of public discussion of the above occurred recently on vms-list [1], where was
introduced a challenge which surfaces naturally in work like the above:

What would make a good candidate for exhibiting the Voynich Manuscript text's mysterious
strangeness?, in the sense that some aspect of it plainly showing it is strange, despite that it
looks so much like just a written language in some unfamiliar alphabet, could easily be put
before an interested newcomer, and a newcomer not necessarily trained in advanced mathematics,
but nevertheless able to appreciate straightforward arithmetic in relation to sequences.

Apparently a simple demonstration of the VMS text's strangeness is not so simple to exhibit.
Nevertheless I maintain that that is an interesting problem in its own right, and a worthy
challenge for serious VMS-text researchers to spend some time thinking about.

As the above work has now grown rather voluminous, it is time to begin presenting its beginnings
in an organized fashion, so that better evaluations of it can be made, and those evaluations
can then influence further efforts.

Here in this communication I wish to deal with {2}, introducing its present definitions, in
preparation for subsequently soon communicating to you the tabulated initial analytic results.
The present definitions, subject to improvement, aim to find a balance with respect to
generality, not too much, not to little:

Definitions for n-graph Spectroscopy conducted on a block of text

{2-1} In this context, a group is a set of symbols, generally regarded as having a sequence
order, so that for example in the group "hello" the symbol "h" is the first, and the symbol "o"
is the last in the sequence-set. A group may be known as a "word" if no ambiguity arises as to
what is being meant.

{2-2} A block of groups is a set of groups or words. The sequence of the groups may or may not
enter critically in some analysis, and if it does then it must be made so explicitly. The format
in which a block is analyzed may or may not matter in a particular analysis; formats include the
List, where each line of the list holds only one group, and the series where each line holds
some number of groups. Provided there is no ambiguity, a block of groups may be termed "text".

Let Ng stand for the total number of groups in a block under consideration. Let Ns stand for the
total number of symbols in the block. Let Vg stand for the average group-size:

Vg = Ns/Ng

{2-3} The Alphabet of a block is the set of distinct symbols from which its groups are
constructed. A sub-Alphabet is a sub-set of the Alphabet, and sub-Alphabets may be taken in some
systematic manner. Let Na stand for the size of an Alphabet or sub-Alphabet, that is the number
of their distinct symbols. So for example, with the Alphabet A-Z, its Na = 26 . Let us further
define for a given block under consideration its total symbols to Alphabet ratio, Sa :

Sa = Ns/Na

It is clear that Sa can never be less than 1.0, unless in some advanced analytic sheme we wish
to introduce something along the lines of a possible singularity in text-sequence space, and
even more strange, negative and complex texts. [2]

{2-3a} An example of the systematic division of an Alphabet into sub-Alphabets is an EVOLVING
Alphabet, according to the frequencies-ranks of the Alphabet-symbols in the block under

Let us suppose we have a text block built upon an Alphabet of 20 symbols: s1 - s20 . That is, we
might have s1 = "a", s2 = "A", s3 = "b", and so on.

Let us further suppose that the ranks of the frequencies (counts) of total occurrences of the
symbols in the block are recorded by : R1 - R20, where Rn records the Alphabet-symbol of
nth rank. So we might have:

R1 = s11
R2 = s2
R3 = s19
R20 = s3

That is, s11 is the most frequent symbol, followed by s2="A", and so on, with s3 here being the
least-frequent symbol in the block under consideration.

Now, to create EVOLVING sub-Alphabets according to symbols' frequencies-ranks, to a resolution
of say 5 symbols, we would create these 4 sub-Alphabets:

1st sub-Alphabet: R1,R2,R3,R4,R5

2nd sub-Alphabet: R1,R2,R3,R4,R5,R6,R7,R8,R9,R10

3rd sub-Alphabet: R1,R2,R3,R4,R5,R6,R7,R8,R9,R10,R11,R12,R13,R14,R15

4th sub-Alphabet: R1,R2,R3,R4,R5,R6,R7,R8,R9,R10,R11,R12,R13,R14,R15,R16,R17,R18,R19,R20

Obviously here, the 4th sub-Alphabet is the completely evolved Alphabet.

{2-4} A Vocabulary is a block with special block characteristics. A Vocabulary is a set of
distinct groups; it may have been synthesized, or it may have been distilled from an existing
set or block of groups.

In general, a block will consist of unique plus repeated groups. For example, this list:


consists of 9 groups, 7 of which are unique, and 2 groups which are a repitition of the non-
unique group "is". Hence there are altogether 8 distinct groups in this block - its Vocabulary
comes to 8 groups.

{2-5} Vocabulary Ratio == (the Vocabulary of a block) / Ng

For the list-example of {2-4} the Vocabulary Ratio = 8/9 = 0.889

If a block of Ng groups consists entirely of distinct groups, then its Vocabulary Ratio
is : (Ng/Ng) = 1.0 .

In contrast to this, a block consisting entirely of Ng identical groups has a Vocabulary Ratio =
1/Ng . As can be seen, a somewhat ambiguous special case arises when Ng=1.

{2-6} For analytic purposes, blocks and Vocabularies, rather than remaining fixed, may be
treated as variables as to one or more of their characteristics, thus permitting block dynamics
analysis. One obvious such characteristic is the size of a block or Vocabulary, that is the Ng
being considered at any given point in the analysis.

Taking up again the example of sub-Alphabets from {2-3a}, we can form 4 sub-blocks, each with
their own Ng, corresponding to the 4 sub-Alphabets: the first sub-block will contain only those
block-groups which are constructed ONLY from the first sub-Alphabet. And so on. And of course
the final sub-block will be the same as the complete source-block. So, Ng will vary as we
consider these sub-blocks each in turn.

And so now we have, to a resolution ultimately determined by the sub-Alphabets, the original
block transformed into a variable, which permits studying the block dynamically, as we shift our
analytic focus around the sub-blocks. If we keep our analytic focus linear from the smallest
sub-Alphabet/block up to the final full size, then we can regard that as one view of the
EVOLUTION of the block - that is, how, in our abstract analytic space, the block evolves along a
path in that space determined by what in the end will be frequencies-ranks of the final complete

Of course other schemes for block and Vocabulary dynamics are possible.

{2-7} During a dynamic analysis of a block or Vocabulary, where its characteristics are 
evaluated stage by stage, the n-graph spectra are of particular interest.

{2-7a} An n-graph spectrum of the block or Vocabulary is the organized set of n-graph amplitude
versus n-graph. Here the:

n-graph amplitude = count of occurrences of the n-graph (i.e. its "frequency") in the groups of
the block or Vocabulary, counted according to some rule such as:

that a given n-graph is counted only once per group even if it appears to occur in the group
more than once.

{2-7b} If an n-graph of the possible n-graph spectrum has at least one occurrence in the block
or Vocabulary, that is its amplitude is at least = 1, then it is an "active" n-graph.

{2-7c} The full n-graph spectrum, akin to "bandwidth", is determined by the Alphabet, or sub-
Alphabet which is governing the particular analysis. If the Alphabet consists of Na symbols,
then the complete n=2 di-graph spectrum would have Na^2 spectral-lines, each of which may or may
not be of zero-amplitude depending on the case.

{2-7d} n-graph spectrum bandwidth == the maximum number of n-graphs possible upon the Alphabet
governing a particular case.

For example, if the total Alphabet is Na=4 : a,b,c,d

then the digraph spectrum bandwidth == Bd = 4^2 = 16

and the digraph spectrum components, to be evaluated for their amplitudes, are:


At this point the ordering of these digraph spectrum components, as shown, is arbitrary. If the
block under consideration with this Na=4 Alphabet contained exactly 1 occurrence of the digraph
"bb", say in the group "Hubble", then the amplitude of the bb spectral component would be 1 :

bb : 1

Likewise, if in the block there was no occurrence of the digraph "da" then:

da : 0

and so on. Of course during a dynamic analysis based on dynamic sub-Alphabets, the digraph or
n-graph spectrum would be changing not just as regards spectral line amplitudes, but also
spectrum bandwidth. Alternately, the dynamic analysis could employ the maximum possible
bandwidth, that being the bandwidth of the complete Alphabet, throughout; in that choice the
bandwidth would be held constant.

{2-7e} n-graph spectrum density == ( no. of active n-graphs ) / ( n-graph spectrum bandwidth )

{2-8} An n-prime symbol of the block or Vocabulary is one of its n most frequently occurring
symbols. Hence the  di-prime  symbols are the first-most and second-most frequent symbols. The 
tri-prime symbols are the 1st-most, 2nd-most, and 3rd-most frequent symbols appearing in the
block or Vocabulary. And so on.

Unless explicitly stated otherwise, we shall take "prime symbols" to mean the tri-prime symbols,
that is the top 3 most-frequent symbols. To gauge their weight among the total symbols Ns, let
us define:

pF == prime symbols frequency ratio = (total of prime-symbols' frequencies) / Ns

In a dynamic analysis, say one proceeding upon changing sub-Alphabets, the prime symbols
identities may change, before stabilizing, if (in general) they stabilize at all.

Using again the nomenclature of {2-3a} : whereas the sub-Alphabets were created according to the
frequency-ranks of the complete Alphabet, where R1, R2, and R3 are the 1st, 2nd, and 3rd most-
frequent symbols of the complete block, nevertheless the identities of the R1, R2, and R3 prime
symbols may be quite different in the sub-blocks.

It is of interest to track the identities of the prime-symbols as the dynamic analysis proceeds.

For example, in a linear evolution analysis we would be interested to see if the R1, R2, R3
identities stabilize early on (smaller sub-Alphabets), stabilize late (approaching the complete
Alphabet and complete block condition), or oscillate, or give no particular indication of
stabilization. To give an indication of stabilization we demand that R1, R2, R3 have the exact
same identities for at least in the complete block, plus in the immediately smaller-to-it sub-
block. If their identities were not these same ones in any of the other sub-blocks, then in the
evolution scenario we could view this as a late stabilization.

If we see oscillation, that is if R1, R2, R3 take on the same prime-symbols' identities here and
there with some sub-blocks and sub-Alphabets, then we are motivated to regard the other symbols
in those sub-Alphabets as having some special relationship to the prime symbols in the context
of the block's groups. That is to say, those other symbols bring groups into the picture which
periodically primize a certain set of symbols as R1, R2, and R3.

{2-9} A prime permutation n-graph is an n-graph constructed from, and only from n-prime symbols.
The complete set of possible prime permutation n-graphs may be regarded as a special sub-
spectrum of the overall n-graphs spectrum. Subject to a possible variation to be discussed in a
moment, there are 4 possible prime permutation di-graphs, and 6 possible prime permutation tri-
graphs with Alphabets of at least 2, and 3 distinct symbols.

So for example, if in any particular case the 3 most frequent / highest-ranking symbols in order are:  I, B, M

then the prime permutation digraphs set is:


and the prime permutation trigraphs set is:


with their order, as just listed, conveniently following the ranks.

{2-10} The definition of the "prime permutation n-graphs frequencies ratio", PPnFR, is motivated
so as to obtain an idea of the relative weight of presence of the pp-ngraphs in the block under
consideration; of course it depends on the general counting-rule mentioned in {2-7a}.

Definition of PPnFR is still subject to some thinking, but currently the attempt is to force the
ratio to always fall only into the range 0 - 1.0 , so that a value of zero would indicate that
the block under consideration is totally devoid of pp-ngraphs, and a value of 1.0 would indicate
the block to be all groups entirely constructed of pp-ngraphs - saturated with them.

Subject to further refinement then, let us presently define as follows for the first two cases:

{2-10a} an n-graph is counted exactly once regardless of how many times once-or-more that it
appears to occur in a group.

{2-10b} PP2FR == ( 2 x total of all pp-digraphs counts ) / Ns

{2-10c} PP3FR == ( 3 x total of all pp-trigraphs counts ) / Ns

Let us see some examples. Suppose we take the examples from {2-9} to be the actual blocks of
groups under consideration. In the case of the first block from {2-9} the pp-digraph counts are:

II : 1
IB : 1
BI : 1
BB : 1

and the total number of symbols Ns = 8

PP2FR = ( 2 x [ 1+1+1+1 ] ) / 8 = 8/8 = 1.0

For the pp-trigraphs example:

IBM : 1
IMB : 1
BMI : 1
BIM : 1
MIB : 1
MBI : 1

the total number of symbols Ns = 18

PP3FR = ( 3 x [ 1+1+1+1+1+1 ] ) / 18 = 18/18 = 1.0

Lets now de-saturate the last example by weakening the weight of the tri-graphs with a new


The prime symbols are still: I,B,M and the pp-trigraphs counts also remain the same:

IBM   : 1
IMBK  : 1
BMI   : 1
BIMuX : 1
MIB   : 1
MBI   : 1

but now Ns=21 and:

PP3FR = ( 3 x [ 1+1+1+1+1+1 ] ) / 21 = 18/21 = 0.857

A moment ago I mentioned in {2-9} a variation we must consider. It's possible, especially with
very small blocks, to have multiple equal-frequency prime symbols, making for equally valid
multiple prime permutation sets. This condition, if it is there, is always immediately apparent
from the block's ordinary symbols-frequencies table. The easiest way to see how an alternate
prime permutations n-graph set could arise is if we imagine that we again have a block whose
prime symbols are I, B, M, but whereas I is unequivocally the most-frequent symbol, B and M are
equally frequent. Hence there are now two possible prime permutation digraph sets:




This complicates matters somewhat, but also makes things more interesting: we could view this
situation in terms of a bifurcation of the prime permutation digraphs spectrum. Similarly with
the pp-trigraphs.

Aside any other utilities they may possess during text-analysis, the PP2FR and PP3FR numbers may
emerge as useful handy rough general indicators of the detailed construction of a block of text.

In [1] I had expressed the optimism that PP3FR alone may be a candidate for fulfilling the
challenge stated at the beginning above - a simple way to exhibit the strangeness of the VMS
"text". It remains to be seen if that comes about - we'll get a better look at the situation
shortly when I communicate to you the table of initial results of experiments with different
text blocks, including Voynich.

In any case, even though the above definitions for n-graph spectroscopy are pretty
straightforward, they can nevertheless support considerably sophisticated or advanced
experimentation with texts. But taken alone, the PP3FR is easy for a newcomer to appreciate:

In your block of text count up the frequencies of the symbols of the block's Alphabet. Take the
three most frequent symbols and form their 6 permutation trigraphs. Count the frequencies of
those 6 trigraphs among the block's groups, counting each trigraph no more than once per group.
Add those trigraphs frequencies together, and divide that sum by the total of symbols in the
block - that's the PP3FR number. It's not difficult to understand - it's an indicator of the
relative presence-weight, in the block of text, of the possible group-composition chunks made up
of the block's three most common symbols.

Berj / KI3U

[1] vms-list thread:
VMs: Emblematic of the strangeness of the Voynich text; launched by Berj / KI3U 23 JUN 2010.

[2] This type of work does have its wild-imagination moments where even without trying we can
suddenly find ourselves contemplating the possibility of defining a text-tensor and then
proceeding to find solutions for the analog of a Schwarzschild radius in text-sequence space.
Actually, a black hole might not be all that bad an analogy for the infamous Voynich text :).

From: Berj N. Ensanian KI3U
To: Journal of Voynich Studies
Sent: 07-14-2010 4:39:11 PM

J.VS: The Castle of Soave

Dear Colleagues

As more and more photographs from around the world are placed online, it is worthwhile
to check what is out there vis-a-vis items to compare with the Castles and other walled
structures in the VMS's nine-rosettes foldout panels.

Here following I've collected some links to photographs of the Castle of Soave, in the
Veneto region of Italy's province of Veneto (northern Italy). The area is well-known for
its wines, and as we know some VMS illustrations can indeed give the impression that
a bunch of grapes is being depicted in some manner or other. The Soave Castle first
dates from the 10th century:,11.2466955%29&z=15

Berj / KI3U

From: Berj N. Ensanian KI3U
To: Journal of Voynich Studies
Sent: 07-21-2010 6:25:12 PM

J.VS: Initial experimental results: n-graph spectroscopy of texts and Vocabularies

Dear Colleagues

Here I present the initial data results, in a Table format, of the experimental work for which the
definitions and basic outlines were introduced in J.VS communication #344 [1].

Since [1] I have introduced one additional definition, the average of the Prime Permutation Digraphs, and
Trigraphs Frequencies Ratios:

AvPP2+3FR == ( PP2FR + PP3FR ) / 2

and this number is given in the below table, Table 346-1.

The data in the table was computed with various functions of the JVSCGS program mentioned in [1] and I
hope I have entered it in the table without errors. The JVSCGS program as you know has thus been getting
its pre-release workout, and some utilities have been added during the progress, and also some of its Help
messages have been improved. I hope JVSCGS in its version V1.0 to be available in the J.VS Library by
the end of this month.

Whereas the amount of data in the table is considerable, in absolute terms it is nevertheless tiny and
just an initial gathering of a base upon which to build and to compare with - I believe the table does
indicate that efforts to develop more such data, and in more detail, is attractive: we see here only
some elementary applications of the possibilities of n-graph spectroscopy, yet they provide quite a useful
and broad picture of the contributions and effects of certain n-graphs in the compositions of texts and
their Vocabularies.

As we discussed off-J, graphs of the data are desirable, and these would permit developing much more of
the actual spectroscopic aspects, rather than as presently, confining ourselves to simple summary numbers.

The table is divided into sections:

346-1-Ca: Fixed full Vocabulary
346-1-Cb: Variable Vocabulary as a function of sub-Alphabets

Following the table I make some observations on its data.


TABLE 346-1 : Vocabulary and Text Comparisons via n-graph Spectroscopy

(note: this Table was prepared in Notepad, so if on your screen it does not display rectangular, with
perfect rows and columns, then just copy it into Notepad.)

Source Identifiers:

H   : Vocabulary of first 11 pages of a novel; {1-1}
HV1 : H Vigenere-enciphered, key = VOYNICH; {1-1}
HV2 : H Vigenere-enciphered, key = HABILITATIONS; {1-1}
R1  : Rules-based transcriptions of selected VMS-astro groups; {1-2}
R2  : De-anagramming of R1; {1-2}
R3  : Translations of R2; {1-2}
V   : VMS Vocabulary derived from GC's voygroup.txt (all soft-space groups eliminated outright); {1-3}
V1  : V with glyphs-expansions: GC-ccc for GC-d; GC-cc for GC-C; GC-ii for GC-I; {1-3}
GL1 : Schiller's Das Lied von der Glocke (German); eszett (resembles Gk. beta) as single letter; {1-4}
GL2 : Schiller's Das Lied von der Glocke (German); eszett replaced by "ss"; {1-4}
DC  : Dante's Commedia (Italian); {1-5}
L   : Latin Dictionary words; {1-6}
E   : John Evelyn's "ACETARIA  A DISCOURSE OF SALLETS"; {1-7}
C1  : chaotically generated groups specified to approximate V's gross-characteristics; {1-8}
C2a : chaotically generated groups; {1-9}
C2b : C2a inflated by a factor of 2; {1-9}
C3: : chaotically generated groups; {1-10}
M   : Askam's Menta Rubea; {1-11}
157 : VMS f1r-57r from GC's voyn_101.txt, soft-spaces made hard; {1-12}
25R : VMS f25r from GC's voyn_101.txt, soft-spaces made hard; Currier-A; {1-13}
58R : VMS f58r from GC's voyn_101.txt, soft-spaces made hard; Currier-A; {1-14}
79V : VMS f79v from GC's voyn_101.txt, soft-spaces made hard; Currier-B; {1-15}
316 : VMS f103r-116r from GC's voyn_101.txt, soft-spaces made hard; {1-16}
J   : Latin Vulgate Genesis Chapter 1 (all capital letters); {1-17}
DS  : Rev. Drury stutter series; {1-18}

v:I = Vocabulary identifier
t:I = Text identifier

Note: double-entries which are identical except for ps indicate bifurcated pp-digraph spectra

 I     | Na |   Ns  |   Sa   |  Ng  | Gr |  Vg  |  ps |   pF  |   Bd |   Dd  |   2R   |   3R   | (2R+3R)/2
 v:H   | 48 |  6801 |  141.7 | 1165 | 13 | 5.84 | ent | 0.269 | 2304 | 0.187 | 0.0499 | 0.0212 |  0.0356
 v:HV2 | 53 |  6801 |  128.3 | 1165 | 13 | 5.84 | omt | 0.183 | 2809 | 0.323 | 0.0306 | 0.0027 |  0.0167
 v:HV1 | 56 |  6801 |  121.5 | 1165 | 13 | 5.84 | 2vy | 0.154 | 3136 | 0.311 | 0.0224 | 0.0000 |  0.0112
 t:R1  | 19 |   184 |    9.7 |   31 | 4-8| 5.94 | ALE | 0.391 |  361 | 0.244 | 0.1196 | 0.0652 |  0.0924
 t:R1  | 19 |   184 |    9.7 |   31 | 4-8| 5.94 | LAE | 0.391 |  361 | 0.244 | 0.1196 | 0.0652 |  0.0924
 t:R2  | 19 |   184 |    9.7 |   31 | 4-8| 5.94 | ALE | 0.391 |  361 | 0.230 | 0.1522 | 0.0000 |  0.0761
 t:R2  | 19 |   184 |    9.7 |   31 | 4-8| 5.94 | LAE | 0.391 |  361 | 0.230 | 0.1522 | 0.0000 |  0.0761
 t:R3  | 21 |   212 |   10.1 |   31 | 5-9| 6.84 | ALN | 0.420 |  441 | 0.193 | 0.1887 | 0.0849 |  0.1368
346-1-C: EFFECTS OF CHANGES IN THE TRANSCRIPTION OF THE SOURCE    * = the prime-symbols set has stabilized
346-1-Ca:       Fixed full Vocabulary
 v:GL1 | 53 |  6931 |  130.8 | 1120 | 15 | 6.19 | ent | 0.331 | 2809 | 0.184 | 0.1088 | 0.0260 |  0.0674
 v:GL2 | 52 |  6982 |  134.3 | 1120 | 15 | 6.23 | ent | 0.328 | 2704 | 0.188 | 0.1080 | 0.0258 |  0.0669
346-1-Cb:       Variable Vocabulary as a function of sub-Alphabets
 v:V   |  5 |   146 |   29.2 |   46 |  6 | 3.17 | c91 | 0.795 |   25 | 0.720 | 0.3698 | 0.2260 |  0.2979
 v:V   | 10 |  7582 |  758.2 | 1504 | 11 | 5.04 | o91 | 0.404 |  100 | 0.920 | 0.0113 | 0.0079 |  0.0096
 v:V   | 15 | 17266 | 1151.1 | 3325 | 11 | 5.19 | o9c | 0.381 |  225 | 0.804 | 0.0093 | 0.0019 |* 0.0056
 v:V   | 20 | 24782 | 1239.1 | 4772 | 12 | 5.19 | o9c | 0.350 |  400 | 0.715 | 0.0074 | 0.0013 |* 0.0044
 v:V   | 25 | 29458 | 1178.3 | 5746 | 12 | 5.13 | o9c | 0.352 |  625 | 0.643 | 0.0069 | 0.0012 |* 0.0041
 v:V   | 30 | 33077 | 1102.6 | 6455 | 12 | 5.12 | o9c | 0.346 |  900 | 0.583 | 0.0066 | 0.0014 |* 0.0040
 v:V1  |  5 |   250 |   50.0 |   68 |  6 | 3.68 | c91 | 0.824 |   25 | 0.720 | 0.5120 | 0.1920 |  0.3520
 v:V1  | 10 |  9376 |  937.6 | 1796 | 11 | 5.22 | co9 | 0.449 |  100 | 0.920 | 0.1785 | 0.0035 |* 0.0910
 v:V1  | 15 | 20377 | 1358.5 | 3810 | 11 | 5.35 | co9 | 0.430 |  225 | 0.809 | 0.1726 | 0.0024 |* 0.0875
 v:V1  | 20 | 27032 | 1351.6 | 5053 | 12 | 5.35 | co9 | 0.388 |  400 | 0.713 | 0.1487 | 0.0018 |* 0.0753
 v:V1  | 25 | 31979 | 1279.2 | 6060 | 12 | 5.28 | co9 | 0.387 |  625 | 0.642 | 0.1441 | 0.0016 |* 0.0729
 v:V1  | 30 | 35310 | 1177.0 | 6697 | 12 | 5.27 | co9 | 0.383 |  900 | 0.551 | 0.1376 | 0.0017 |* 0.0697
 I     | Na |   Ns  |   Sa   |  Ng  | Gr |  Vg  |  ps |   pF  |   Bd |   Dd  |   2R   |   3R   | (2R+3R)/2
346-1-D: VARIABLE VOCABULARY AS A FUNCTION OF SUB-ALPHABETS       * = the prime-symbols set has stabilized
 v:DC  |  5 |   151 |   30.2 |   50 |  6 | 3.02 | rae | 0.729 |   25 | 0.800 | 0.4238 | 0.0795 |  0.2517
 v:DC  | 10 | 10563 | 1056.3 | 1792 | 12 | 5.90 | tae | 0.364 |  100 | 0.900 | 0.1094 | 0.0139 |  0.0617
 v:DC  | 15 | 44253 | 2950.2 | 6768 | 15 | 6.54 | aei | 0.318 |  225 | 0.800 | 0.0046 | 0.0001 |  0.0024
 v:DC  | 20 | 79932 | 3996.6 | 11712| 16 | 6.83 | aie | 0.319 |  400 | 0.680 | 0.0279 | 0.0002 |* 0.0141
 v:DC  | 25 | 88554 | 3542.2 | 12911| 16 | 6.86 | aie | 0.317 |  625 | 0.534 | 0.0278 | 0.0001 |* 0.0140
 v:DC  | 30 | 92090 | 3069.7 | 13495| 16 | 6.82 | aie | 0.317 |  900 | 0.424 | 0.0281 | 0.0002 |* 0.0142
 v:DC  | 35 | 95061 | 2716.0 | 13962| 16 | 6.81 | aie | 0.317 | 1225 | 0.356 | 0.0282 | 0.0002 |* 0.0142
 v:DC  | 40 | 97012 | 2425.3 | 14297| 16 | 6.79 | aie | 0.316 | 1600 | 0.325 | 0.0280 | 0.0002 |* 0.0141
 v:DC  | 45 | 97560 | 2168.0 | 14402| 16 | 6.77 | aie | 0.315 | 2025 | 0.278 | 0.0279 | 0.0002 |* 0.0141
 v:DC  | 48 | 97582 | 2033.0 | 14407| 16 | 6.77 | aie | 0.315 | 2304 | 0.247 | 0.0279 | 0.0002 |* 0.0141
 v:L   |  5 |    47 |    9.4 |   15 | 2-6| 3.13 | sue | 0.787 |   25 | 0.560 | 0.4681 | 0.0638 |  0.2660
 v:L   | 10 |  3287 |  328.7 |  540 | 13 | 6.09 | sto | 0.346 |  100 | 0.890 | 0.0469 | 0.0100 |  0.0285
 v:L   | 15 | 18826 | 1255.1 | 2708 | 15 | 6.95 | ois | 0.294 |  225 | 0.836 | 0.0291 | 0.0096 |  0.0194
 v:L   | 20 | 29710 | 1485.5 | 4211 | 16 | 7.06 | ioe | 0.293 |  400 | 0.695 | 0.0269 | 0.0003 |* 0.0136
 v:L   | 25 | 31312 | 1252.5 | 4419 | 18 | 7.09 | ioe | 0.292 |  625 | 0.518 | 0.0260 | 0.0003 |* 0.0132
 v:L   | 30 | 31484 | 1049.5 | 4441 | 18 | 7.09 | ioe | 0.291 |  900 | 0.377 | 0.0260 | 0.0004 |* 0.0132
 v:L   | 35 | 31563 |  901.8 | 4450 | 18 | 7.09 | ioe | 0.291 | 1225 | 0.284 | 0.0260 | 0.0004 |* 0.0132
 v:L   | 39 | 31595 |  810.1 | 4454 | 18 | 7.09 | ioe | 0.291 | 1521 | 0.232 | 0.0260 | 0.0004 |* 0.0132
 v:E   |  5 |    52 |   10.4 |   17 |  5 | 3.06 | esi | 0.712 |   25 | 0.560 | 0.4615 | 0.2885 |  0.3750
 v:E   | 10 |   882 |   88.2 |  191 | 11 | 4.62 | ste | 0.429 |  100 | 0.850 | 0.1202 | 0.0442 |  0.0822
 v:E   | 15 |  7144 |  476.3 | 1106 | 13 | 6.46 | ens | 0.308 |  225 | 0.831 | 0.0591 | 0.0147 |  0.0369
 v:E   | 20 | 14810 |  740.5 | 2193 | 16 | 6.75 | eir | 0.288 |  400 | 0.748 | 0.0169 | 0.0099 |  0.0134
 v:E   | 25 | 23825 |  953.0 | 3433 | 16 | 6.94 | eir | 0.268 |  625 | 0.600 | 0.0153 | 0.0086 |  0.0120
 v:E   | 30 | 29475 |  982.5 | 4269 | 22 | 6.90 | eis | 0.262 |  900 | 0.506 | 0.0155 | 0.0085 |  0.0120
 v:C1  |  5 |   338 |   67.6 |  109 |  6 | 3.10 | e18 | 0.645 |   25 | 1.000 | 0.2663 | 0.0976 |  0.1820
 v:C1  | 10 |  1938 |  193.8 |  569 |  8 | 3.41 | 8ea | 0.383 |  100 | 0.990 | 0.0970 | 0.0186 |  0.1156
 v:C1  | 15 |  4373 |  291.5 | 1214 | 12 | 3.60 | h8e | 0.331 |  225 | 0.982 | 0.0604 | 0.0117 |* 0.0361
 v:C1  | 20 |  8996 |  449.8 | 2262 | 12 | 3.98 | h8e | 0.293 |  400 | 1.000 | 0.0478 | 0.0100 |* 0.0289
 v:C1  | 25 | 17412 |  696.5 | 3883 | 12 | 4.48 | h8e | 0.267 |  625 | 0.998 | 0.0404 | 0.0069 |* 0.0237
 v:C1  | 30 | 33173 | 1105.8 | 6455 | 12 | 5.14 | h8e | 0.242 |  900 | 1.000 | 0.0344 | 0.0051 |* 0.0198
 I     | Na |   Ns  |   Sa   |  Ng  | Gr |  Vg  |  ps |   pF  |   Bd |   Dd  |   2R   |   3R   | (2R+3R)/2
346-1-E: COMPLETE SOURCE-TEXT VS ITS VOCABULARY             change:  ^ = upward   _ = ~none   v = downward
 t:M   | 34 |  2403 |   70.7 |  594 | 10 | 4.05 | etn | 0.357 | 1156 | 0.204 | 0.0683 | 0.0462 |^ 0.0573
 v:M   | 34 |  1220 |   35.9 |  235 | 10 | 5.19 | ets | 0.335 | 1156 | 0.204 | 0.0771 | 0.0148 |  0.0460
 t:J   | 22 |  2776 |  126.2 |  537 | 17 | 5.17 | ETI | 0.363 |  484 | 0.391 | 0.1023 | 0.0043 |^ 0.0533
 v:J   | 22 |  1356 |   61.6 |  211 | 17 | 6.43 | EIA | 0.334 |  484 | 0.391 | 0.0207 | 0.0000 |  0.0104
 t:DS  | 22 |   639 |   29.1 |  158 | 13 | 4.04 | TEA | 0.316 |  484 | 0.296 | 0.0344 | 0.0000 |^ 0.0172
 v:DS  | 22 |   399 |   18.1 |   88 | 13 | 4.53 | ETA | 0.288 |  484 | 0.296 | 0.0251 | 0.0000 |  0.0126
 t:E   | 63 | 144589| 2295.1 | 31478| 22 | 4.59 | eta | 0.286 | 3969 | 0.207 | 0.0394 | 0.0156 |^ 0.0275
 v:E   | 63 |  42856|  680.3 | 6333 | 22 | 6.77 | eis | 0.260 | 3969 | 0.207 | 0.0152 | 0.0088 |  0.0120
 t:C3  | 38 |  96004| 2526.4 | 24395| 10 | 3.94 | eta | 0.168 | 1444 | 1.000 | 0.0130 | 0.0006 |  0.0068
 v:C3  | 38 |  88474| 2328.3 | 19949| 10 | 4.44 | eta | 0.163 | 1444 | 1.000 | 0.0130 | 0.0007 |_ 0.0069
 t:C2b | 38 |  1078 |   28.4 |  200 | 12 | 5.39 | &tn | 0.161 | 1444 | 0.242 | 0.0186 | 0.0000 |_ 0.0093
 v:C2a | 38 |   539 |   14.2 |  100 | 12 | 5.39 | &tn | 0.161 | 1444 | 0.242 | 0.0186 | 0.0000 |  0.0093
 t:157 | 149| 37938 |  254.6 | 10261| 10 | 3.70 | o91 | 0.384 | 22201| 0.036 | 0.0047 | 0.0048 |v 0.0048
 v:157 | 149| 16258 |  109.1 | 3518 | 10 | 4.62 | o91 | 0.367 | 22201| 0.036 | 0.0085 | 0.0055 |  0.0070
 t:25R | 26 |   180 |    6.9 |   49 |  7 | 3.67 | oa1 | 0.372 |  676 | 0.090 | 0.0000 | 0.0000 |v 0.0000
 v:25R | 26 |   161 |    6.2 |   42 |  7 | 3.83 | o19 | 0.385 |  676 | 0.090 | 0.0745 | 0.0000 |  0.0373
 t:58R | 50 |  1651 |   33.0 |  399 |  9 | 4.14 | aoe | 0.442 | 2500 | 0.072 | 0.0049 | 0.0454 |v 0.0252
 v:58R | 50 |  1333 |   26.7 |  292 |  9 | 4.57 | oae | 0.421 | 2500 | 0.072 | 0.0060 | 0.0540 |  0.0300
 t:79V | 41 |  1534 |   37.4 |  383 |  8 | 4.01 | o9c | 0.393 | 1681 | 0.092 | 0.0013 | 0.0000 |v 0.0007
 v:79V | 41 |   972 |   23.7 |  226 |  8 | 4.30 | 9oc | 0.406 | 1681 | 0.092 | 0.0021 | 0.0000 |  0.0011
 t:316 | 100| 46287 |  462.9 | 11816| 10 | 3.92 | oa9 | 0.344 | 10000| 0.069 | 0.0031 | 0.0001 |v 0.0016
 v:316 | 100| 17078 |  170.8 | 3572 | 10 | 4.78 | oc9 | 0.331 | 10000| 0.069 | 0.0817 | 0.0012 |  0.0415
 I     | Na |   Ns  |   Sa   |  Ng  | Gr |  Vg  |  ps |   pF  |   Bd |   Dd  |   2R   |   3R   | (2R+3R)/2

Na = size of Vocabulary's or Text's Alphabet or sub-Alphabet
Ns = total number of symbols in the Vocabulary or Text
Sa = total symbols to Alphabet ratio = Ns/Na
Ng = total number of groups or words in the Vocabulary or Text
Gr = range of group-sizes in the Vocabulary or Text from 1 (unless other shown) to ...
Vg = average group-size = Ns/Ng
ps = prime symbols in rank order
pF = prime symbols frequency ratio = (total of prime symbols' freqs.)/Ns
Bd = digraph spectrum bandwidth
Dd = digraph spectrum density
2R = prime permutation digraphs frequencies ratio  PP2FR
3R = prime permutation trigraphs frequencies ratio PP3FR

{1-1} H and HV1 and HV2 were prepared by Dennis Fedak / N3ZCK. The Vocabulary list H is distilled from the
first 11 pages of the romance novel "Finally Home" by Dawn Michele Werner, from this url:
HV1 is a Vigenere enciphering of H, where each list-word in H is enciphered as if it were a complete
plaintext message, using the key-word "VOYNICH". HV2 is similarly enciphered, using the key "HABILITATIONS".

{1-2} Robert Teague's ongoing VMS astro-section Labels work, which is periodically published (e.g. see the
J.VS Library). R1 is data provided by Robert from his private research - it is a list of VMS labels
transformed by one system of letters-values, and within the context of the analysis here it is taken as a
list of system-transcribed VMS astro-section labels. R2 is de-anagrammings of R1, and R3 is the trans-
lations of R2, for example R2-list-group ALDEBRAN is translated to R3-list-group ALDEBARAN.

{1-3} Glen Claston's (GC) VMS transcriptions are available here at Voynich Central:
GC's 2006-released Voynich groups Concordance, voygroup.txt, was used as the source for V - the Vocabulary
of the VMS was extracted from this file with the condition that all groups connected with "soft-spaces"
were discarded outright. V1 is V with the following glyphs-expansions:
GC-ccc for GC-d; GC-cc for GC-C; GC-ii for GC-I

{1-4} GL1 is the German text of Schiller's "Das Lied von der Glocke" take from this url:
GL1 retains the Latin motto at the beginning of the poem after removal of the surrounding brackets and
asterisk. GL1 retains the single-letter eszett (resembles the Greek beta). GL2 is GL1 with one letter-
expansion: the eszett is replaced with "ss".

{1-5} DC is derived from the August, 1997 Etext #1012, of the Project Gutenberg transcription of Dante's
Commedia, in Italian, dowloaded from this url:
I had quite some difficulty deciding how to handle extracting the Vocabulary from this work, on account
of the orthography and my unfamiliarity with properly handling written Italian text. Therefore the data
in the table for DC should be taken as only very roughly accurate.

{1-6} L is derived from the Latin Dictionary and Grammar Aid provided online by the University of Notre Dame
at this url:
The download was done DEC 2007. The L file was prepared from the download by extracting only the provided
Latin Vocabulary words.

{1-7} E is derived from April 2005 EBook 15517.txt of the Project Gutenberg transcription of John Evelyn's
(1620-1700) 1st edition (1699) "ACETARIA  A DISCOURSE OF SALLETS", downloaded from this url:
E is based on the basic body text of Evelyn's work: omitting the index, and coming to 31478 words. The full
Alphabet is 63 symbols, full Vocabulary 6333 words, Vocabulary Ratio = 0.201

{1-8} C1 is based on a JVSCGS chaotically generated block of pseudo-random groups, generated so as to
approximate the gross block characteristics of the 30-symbols sub-Vocabulary of V.

{1-9} C2a is a JVSCGS chaotically generated block of pseudo-random groups, and C2b is an inflation of C2a
by a factor of 2: a copy of C2a was appended to itself. In the table C2b with 200 groups is taken as the
"text, while C2a with 100 groups is C2b's Vocabulary. Therefore C2b's "Vocabulary Ratio" = 100/200 = 0.5
The purpose of C2a and C2b is to show which numbers in the table remain constant with uniform inflation of
the groups.

{1-10} C3 is a JVSCGS chaotically generated block of pseudo-random groups, being a combining of two
separately generated sub-blocks. The two sub-blocks, 12000 and 12395 groups each, were each generated with
identical specifications for gross block-characteristics, EXCEPT that the first sub-block was generated
with group-sizes 1-10, while the second block was generated with group sizes only 3-10. This was done so
that the final combination block would have a group-sizes frequencies distribution bulging closer to the
middle of the 1-10 range, instead of toward the low end. The combined block's group-size vs frequency is:

 group-size : frequency

 01 :  2494
 02 :  1981
 03 :  7348
 04 :  7658
 05 :  880
 06 :  791
 07 :  775
 08 :  767
 09 :  845
 10 :  856

The more-in-the-middle bulge provides more chances for digraphs and trigraphs of interest to arise.

{1-11} "A litle Herball of the properties of Herbes, ..... ", by Anthony Askham, transcribed by
Glen Claston from both the 1550 and 1553 editions:
The Menta Rubea portion of Askham's work for M in the above table was taken directly from {3} in J.VS
communication # 207, by Berj / KI3U, (Vol. II, 27 JUL 2008):
J.VS: NvP Sequence Spectra Topographic comparisons: Voynich folio f111r text vs Askham's Menta Rubea ;
with of course the line numbers and groups counts at left margin removed from {3}.

{1-12} The GC-transcribed text of Voynich Manuscript folios f1r-57r, but for f1r being all botanical
folios, taken as one block; for the work here the soft spaces in the transcription were first all made
into hard spaces.

{1-13} The GC-transcribed text of Voynich Manuscript folio 25r; for the work in this table the soft spaces
in the transcription were first all made into hard spaces. The f25r botanical illustration has been
suggested by Dana Scott as depicting the Mint herb: vms-list post by Dana Scott: Sun, 04 Nov 2001 23:06:47 -0800 : VMS -- Botany (f25r) Mint

{1-14} The GC-transcribed text of Voynich Manuscript folio 58r, it is along with 58v a star-folio, somewhat
akin to the star-pages (f103r-116r) which end the VMS; for the work in this table the soft spaces
in the transcription were first all made into hard spaces.

{1-15} The GC-transcribed text of Voynich Manuscript folio 79v; for the work in this table the soft spaces
in the transcription were first all made into hard spaces. This VMS folio is in the "biological" section
of the manuscript, and is distinguished from all other folios in that its illustrations include the
clearest example of apparently Christian symbolism: at the top a VMS "nymph" is see holding in her out-
stretched hand a cross and with it "blessing" the folio's text.

{1-16} The GC-transcribed text of Voynich Manuscript star folios f103r-116r, taken as one block; for work
in this table the soft spaces in the transcription were first all made into hard spaces.

{1-17} Latin Vulgate Genesis Chapter 1, downloaded from DRBO.ORG :
For work in this table the verse identifiers were removed, and the text was rectified to all capital letters.

{1-18} The Billy Graham Center has a transcript of a 1997 interview by Robert Shuster of the Rev. William A.
Drury that was prepared with great care and records Drury's stammering / stuttering: Billy Graham Center,
Archives, Collection 492 - Rev. William A. Drury. T3 Transcript. Downloaded 6 MAY 2008 from this url:
The DS block in this table's work is a small portion of the complete Drury transcript - it is the portion
given in footnote [14] of J.VS communication # 203, by Berj / KI3U, (Vol. II, 30 JUN 2008):
J.VS: Reference Data for Analysis of Signal Sequence Series ;
The "_" group-separators were removed.


* ++++++++++++ Some observations on the data of Table 346-1 ++++++++++++ *


The longest word in the source plaintext is 13 characters. It would appear from the results that the
13-characters-long key "HABILITATIONS" is not as effective as the shorter key "VOYNICH" in scrambling
the source plaintext - the VOYNICH-key Vigenere output (HV1) shows lower PP2FR and PP3FR numbers.
Unlike the shorter VOYNICH, the HABILITATIONS key's composition has repititions of characters.


In these examples the process is opposite that of 346-1-A and the attempt is to decipher the source.
Going the major step from the R1 source (selected labels from the VMS astro-section) to R2, we see a
zero-ing of the PP3FR, but the PP2FR rises - as if digraphs are being "recovered". The AvPP2+3FR is
lower in R2 than in R1. Both R1 and R2 have bifurcated pp-digraph spectra. "Translating" R2 into R3,
the AvPP2+3FR rises again, and while the tri-prime symbols set changes, the di-prime symbols prevail.

346-1-Ca: Fixed full Vocabulary

In some sense here the transcription is the indpendent variable. The only difference between the GL1
and GL2 source-Vocabularies is that the single-character eszett of GL1 has been made the double "ss"
in GL2. With a fairly hefty Ns, greater than 6900, plus the fact that the eszett is not a prime-symbol,
it is perhaps not surprising that varying the transcription in going from GL1 to GL2, the summary
spectroscopic changes are quite small - just a slight decrease in PP2FR, PP3FR, and AvPP2+3FR. So then
the net effect here of the transcription substitution of ss for eszett is a slight diluting of the
relative presence-weight of the prime permutation digraphs and trigraphs.

346-1-Cb: Variable Vocabulary as a function of sub-Alphabets

These table entries, although placed together, can be studied for two different effects: the effect of
changes in transcription as in the preceeding case, and the effect arising from turning Vocabulary into
a variable in some systematic manner.

The method of turning the VMS Vocabulary into a variable here is exactly the one described in {2-3a}
and {2-6} of [1] : in our abstract analytic space, the source Vocabulary, as a variable, evolves along
a path in that space determined by what in the end will be the frequencies-ranks of the final complete
Alphabet. In the table the resolution of the evolution-path is 5 Alphabet-symbols, and the path is
followed up to 30 Alphabet-symbols (there are many more symbols in the complete source, and the path
could be studied all the way to its end, but the vast majority of the Voynich Vocabulary is covered by
its 30 most-frequent symbols).

In contrast with the GL2 vs GL1 transcription changes, here V1's transcription vs V is considerably more:
the glyphs-expansions: GC-ccc for GC-d, GC-cc for GC-C, GC-ii for GC-I, are used in V1, and moreover GC-c
is almost always a tri-prime symbol in both V and V1. As we can see in section 346-1-Cb of the table,
whereas GC-c is always a tri-prime symbol with V1, hardly surprising considering the transcription in
effect, with V it does temporarily vanish as a tri-prime, as the Vocabulary variable, itself a function
of the ultimate symbols' frequencies ranks, transitions though the top-ten ranks, Na = 10. This is surely
related to another obvious difference between the evolutions of V and V1: the tri-prime symbols set of V1
stabilizes earlier than does that of V.

In V, GC-9 is stable throughout as the prime symbol of 2nd rank. In V1, GC-c is stable throughout as the
prime symbol of 1st rank, and GC-9, although it starts out in 2nd rank, quickly drops to 3rd rank with the
stabilization of the tri-prime set. Now, if we did not know that V and V1 are transcription variations of
one another, then we could not make much of this. But knowing it, we see a logical or perhaps even
expected pattern: in both transcripts GC-9 immediately follows GC-o in rank, and the change in the
transcription has only moved GC-c from 3rd to 1st rank, once stabilization is entered.

With both V and V1, prior to stabilization, only GC-1 is a temporary member of the tri-primes sets.

Between the evolutions of V and V1 the digraph spectrum densities, Dd, do not become especially different
until Na = 30 is reached, and V1's exceeds V's - again not surprising considering the transcription in
effect, and actually one might wonder why it took so long.

But it is with the PP2FR, PP3FR, and AvPP2+3FR numbers that we can immediately tell that there must be some
major differences in the composition of substantial portions of the groups of the respective sources.
Whereas the glyphs expansions raise the PP3FR in V1 just a bit, they dramatically increase the PP2FR -
the glyphs expansions make V1's digraph spectrum come alive. This is expected since the glyphs expansions
are mostly x2 multiplications of the symbols. But in the V1 source the ultimate frequency rank of GC-i is
only down at 21st, so it would be indicated to see just how much the GC-ii subsitution for GC-I is
contributing to the PP2FR number.

Finally, along the variable Vocabulary path, lets have a look at the rate-of-change of PP2FR. Those numbers
are not entered in the table, but we can quickly calculate a representation of them by forming the
successive PP2FR ratios:

for V: 32.73, 1.22, 1.26, 1.07, 1.05
for V1: 2.92, 1.03, 1.16, 1.03, 1.05

Viewing these numbers as digitized curves, both of these pulse-decay-type curves have a slight bulge in the
middle, with V's being bigger and broader, and V1's more sharply peaked and symmetric. Apparently this means
that the evolving Vocabulary, dependent as it is on the Alphabet, acquires a bunch of digraph-rich groups in
the neighborhood of the Alphabet = 20 symbols mark, which in experienced VMS text work immediately calls to
mind the well-known observation that most of the VMS text is dominated by around 17 or so glyphs.

So even though V1's transcription is non-trivially different from V's, the signal of the digraph-rich
bunch is evident in both transcriptions. And what is rather intriguing about this is that the two
respective transcription bulges can almost give the impression of being Fourier transforms of one another -
this may be worth looking into for texts in general.


Continuing the same kind of evolution of Vocabulary analysis, here we compare the Vocabularies of Dante's
Commedia (in Italian) DC, a Latin Dictionary L, a 1699 English work by John Evelyn E, and a chaotically
generated block of pseudo-random groups approximating V in its gross characteristics C1. As noted in the
table footnote {1-5} the DC data should be considered only roughly accurate.

Looking at the tri-prime sets, C1 stabilizes fairly early, about like V and V1. DC and L stabilize somewhat
later, and E doesn't stabilize at all, instead exhibiting indications of oscillation discussed in {2-8} of
[1]. Prior to stabilization, two symbols temporarily appear in the prime-sets of DC, later to vanish from
the stabilized set. With L there are three such temporaries. We recall from a moment ago that with V and V1
there was just one temporary prime symbol. With C1 there are two temporaries. Therefore for experiments
seeking to match Latin with the VMS, it may be of interest to chaotically assemble a Latin Vocabulary so
that its dynamic prime-symbols sets mimick this situation with the VMS Vocabulary.

Especially with DC and L we see that the prime symbols frequency ratio, pF, fairly stabilizes early on,
hardly changing at all once the tri-primes have stabilized.

The digraph spectrum density, Dd, remains at or nearly 1.0 with C1 - I take this to mean that the C1 block
of pseudo-random groups does indeed exhibit randomness: during its generating any digraph spectral-line
has as much chance as any other to become active. Taking the 30-Alphabet-symbols mark on the variable
Vocabulary path, by that point the Voynich V and v1 have a more active digraph spectrum, their Dd's being
0.583 and 0.551 respectively, than all the others except for C1.

The table does not provide the ratio: ( PP2FR / PP3FR ), but we can quickly calculate it from the table
data. This ratio should indicate the relative influences of prime digraphs vs trigraphs. Lets obtain these
numbers for the 30-Alphabet-symbols mark and see what they come to:

for V: 4.71
for V1: 80.94
for DC: 140.5
for L: 65
for E: 1.82
for C1: 6.75

Since the goal of generating C1 had been to approximate V's gross characteristics, we see it was fairly
successful in this measure. The tremendous difference between the V and V1 numbers is really no surprise
since the transcription changes in V1 favored digraph production. Evidently the English E and the Italian
DC are very far apart in this measure. In the side-thoughts which occur in work of this kind, we ponder an
interesting problem: constructing for several different languages sub-Vocabularies of their total known
Vocabularies, which have matching n-graph spectral characteristics, and then investigating to what extent
those thus-constrained different-language Vocabularies can effectively communicate the same discourses.

Finally, at the 30-Alphabet-symbols mark the AvPP2+3FR numbers are really interesting: V is the lowest
at 0.0040 and V1 is the highest at 0.0697 . The others all fall into the 0.01-something range, even the
chaotic C1. Aside from this being ponderable as a possible indicator of the oddballness of the VMS, it
does highlight one of the old problems dogging VMS text research: what is the correct transcription?
Although the transcription differences between V and V1 are non-trivial, they are far from the extreme
they could be, especially when we think about the problem of the socalled intruding gallows glyphs -
just what is the correct way to transcribe them? Whereas some washout of these problems occurs when big
blocks are analyzed in summary-fashion, these problems magnify terrifically at the paragraph-block level.


In this section of the table the entries for chaotically generated Vocabulary and text serve as a
convenient reference. We see that as the chaotic Vocabularies morph into text, the average of their
prime permutations digraph and trigraph frequencies ratios, the AvPP2+3FR numbers, remain unchanged.

With the four normal language examples the AvPP2+3FR increases as the Vocabulary morphs into text.

But the opposite happens with VMS text, both for small and large blocks of groups - the AvPP2+3FR drop
as the VMS Vocabularies morph into VMS text. So then, within the confines of the data in the table, the
Voynich Manuscript writing behaves like a kind of "anti-text". It remains to be discovered, with the
accumulation of much more data, how unusual this remains, and in particular if the AvPP2+3FR measure is
able to serve as a useful simple gauge for demonstrating the strangeness of the Voynich text: on the
folios it gives the impression of some sort of writing, albeit in unfamiliar script, yet it in some sense
it is an anti-text, this anti-text characteristic being even more pronounced in the last part of the book
compared with the early part.

* ++++++++++++ Some concluding comments ++++++++++++ *

The focus of the table data is the prime permutation digraphs and trigraphs. It is also interesting to
investigate non-prime permutation n-graphs, that is permuation n-graph sets comprised of non-prime or mixed
prime and non-prime Alphabet symbols. For instance, in the table we have seen that in evolving Vocabulary
analyses, sometimes there are temporary di-prime and tri-prime symbols: what are their effects in general?

Lets have a start on this with a quick look. Above in 346-1-D: VARIABLE VOCABULARY AS A FUNCTION OF SUB-
ALPHABETS, we found that With L there are three temporary primes; they are: s, u, t. From the table we
see that it is not until the 15-symbols mark that the ultimate tri-prime symbols begin to dominate the
prime set, requiring until the 20-symbols mark to complete their trio. At the 5 and 10-symbols marks it is  s, u, t,
which dominate the evolving prime-symbols set.

In the complete L Vocabulary of 39 Alphabet-symbols the frequencies and ranks of s, u, t, are:

symbol : frequency : frequency-rank

u : 2745 : 4
s : 2616 : 5
t : 2025 : 8

Let us form their permutation digraphs and trigraphs sets, and obtain the variable-Vocabulary P2FR, P3FR,
and AvP2+3FR (as we shall denote them) for those, corresponding to the resolution in 346-1-D:

Na : P2FR : P3FR : AvP2+3FR

05 : 0.4681 : 0.0000 : 0.2341
10 : 0.1363 : 0.0685 : 0.1024
15 : 0.0923 : 0.0402 : 0.0663
20 : 0.0852 : 0.0338 : 0.0595
25 : 0.0834 : 0.0332 : 0.0583
30 : 0.0832 : 0.0331 : 0.0582
35 : 0.0831 : 0.0330 : 0.0581
39 : 0.0831 : 0.0330 : 0.0581

We see the P3FR spiking at the 10-symbols mark - this is in the region of the Vocabulary's evolution where
the ultimate tri-prime symbols (i, o, e) are not yet dominant. Otherwise, the numbers trend much like the
PP2FR, PP3FR, and AvPP2+3FR numbers in 346-1-D computed for the ultimate tri-prime symbols, but their
magnitudes are considerably greater: as the L Vocabulary evolves, the s, u, t, symbols contribute more
heavily to the digraph and trigraph spectra than do the ultimate tri-prime symbols.

The PP2FR, PP3FR, and AvPP2+3FR, and as well any particular P2FR, P3FR, and AvP2+3FR, are relatively easy
to compute, and most especially they are easy to understand with just moderate mathematical skills, but
nevertheless as we have seen above, they seem to provide a useful perspective on the detailed constructions
of Vocabularies and texts based on them.

Berj / KI3U

[1] Journal of Voynich Studies communication #344 (Vol. IV, 14 JUL 2010), by Berj / KI3U :
J.VS: Definitions for n-graph spectroscopy

From: Berj N. Ensanian KI3U
To: Journal of Voynich Studies
Sent: 07-26-2010 12:45:22 PM

J.VS: The JVSCGS Chaotic Groups Generator and Utilities computer program

Dear Colleagues

I have sent to our J.VS Librarian Greg Stachowski the version V1.0 of the JVSCGS computer program,
as J.VS Library Deposit # 30-1-2010-07-25, which will hopefully be online for download in a day or so:

JVSCGS is the computer program which was used to obtain the data in Table 346-1 presented in
J.VS communication #346 [1]. Additional mention of JVSCGS is in [2,3].

JVSCGS is intended as a tool for general text or numbers sequences research, in particular with the Voynich
manuscript text, and contains three major components:

1.) a generator {G} of pseudo-random groups sequences built upon the interaction of 4 chaotic functions -
this is the core of JVSCGS. The generator is steerable so that output sequences which appear promising as to
their affinity with texts under investigation, may be steered for better affinity. A variety of functions
including symbols and n-graphs mappings are available to tailor the generator output. Many characteristics
measuring, and alternately characterizing the generated output for analytic purposes, are available for
screen-display and disk saving.

2.) a simpler chaotic sequences generator [R], which in addition to serving as a generator, is also useful
for helping gain experience with the much more sophisticated generator {G}, and try out limited tests to
ascertain interesting parameters for experiments with {G}.

Both 1.) and 2.) are first and foremost chaotic generators of integers, and these integer sequences are
always available - any interesting pattern in generated output must ultimately be investigated among the
generating integers.

3.) Utilities |V| for preparing and analyzing data files, both generated by JVSCGS, and external like VMS
transcription files available elsewhere.

JVSCGS was compiled for Windows console, and was developed on a Windows 32 Vista computer.
The program should be very helpful to newcomers to VMS text analysis for one particular reason: experimenting
with the program will save the newcomer time in trying to formulate his / her own computer programming goals
with respect to VMS text analysis - it takes quite a while for a newcomer to become sufficiently familiar
with the VMS text, and crystallize their ideas on how to attack it and what specific attack programs to write.

The above mentioned J.VS Library deposit also has a zipfile of screen-shots of JVSCGS in use - together
with the program's built-in Help messages these will hopefully suffice to work effectively with the program.

It is very useful in the beginning to create some simple very short test text-files to run against JVSCGS's
functions, so as to quickly determine its operations in the minutest arithmetic detail. The program's
single-step function is also very useful for gaining familiarization.

JVSCGS presents practically an astronomical spectrum of generating possibilities when all its resources are
switched in. This is especially so with the symbols and n-graphs mappings. I anticipate that in the future
interesting map files (which can be directly loaded by JVSCGS) will be communicated and made available in
the J.VS Library.

Of course the more complex the task given to the groups generator, the longer the computation will take,
also depending on one's computer's power and how heavy the computer's background load is. Most of the time-
intensive computations I have been doing come in at under a half a minute, a few considerably longer. As you
know from off-J, back in early May during the early stages of the development of JVSCGS, our colleague
Greg Stachowski optimized some of the generating code to make it run faster, and the result was that one test
which had originally needed 16 minutes to compute, computed in mere seconds. To become familiar with the
program's performance on your own computer, work upward in complexity from very simple tasks given to the

The V1.0 program is not perfect, and no doubt can be improved here and there. However its readyness for
release is seen in the Table 346-1 data - all that was computed in a breeze, so to speak, without any
glitches. V1.0 is however sparse in its user-error checking - it expects the user, for example, not to ask
it to access a diskfile which does not exist. In this respect JVSCGS V1.0 should be regarded not as consumer
software, but a dedicated researcher's program.

Berj / KI3U

[1] J. Voynich Studies communication #346 (Vol. IV, 21 JUL 2010):
J.VS: Initial experimental results: n-graph spectroscopy of texts and Vocabularies, by Berj / KI3U.

[2] Journal of Voynich Studies communication #344 (Vol. IV, 14 JUL 2010), by Berj / KI3U.
J.VS: Definitions for n-graph spectroscopy

[3] vms-list thread:
VMs: Emblematic of the strangeness of the Voynich text; launched by Berj / KI3U 23 JUN 2010.

From: Berj N. Ensanian KI3U
To: Journal of Voynich Studies
Sent: 07-30-2010 9:25:08 PM

J.VS: n-graph spectroscopy summary of a Log

Dear Colleagues

Yesterday on vms-list [1] I mentioned to our colleague Richard SantaColoma my intention to run an n-graph
spectroscopy study, a la Table 346-1 [2], on a Log type list. I ran it and here I give the results.

The log is of longwave aeronautical beacons recorded with one particular radio receiver, the 1940's vintage
Learavian RM 402C, and you are familiar with the details from off-J. Relevant to our purposes here the
format of this log is 48 lines of highly condensed / abbreviated data, of which this line is typical:

DDP  :  391 : San Juan Puerto Rico : 1 KW : 8 OCT 2009 eve.; 400-Learav. SSE-NNW; BFD-DX 1766 mi.

All 48 log entries are different, many with more details than the above, but 10 entries represent 5 beacons
recorded twice, once during last winter, and once this summer. The JVSCGS utilities [3] were used to prepare
and analyze the log. To prepare the source lines for analysis, any periods, commas, colons, semicolons,
slashes and hyphens were treated as if they were blank spaces. Therefore the above line became these 20 groups:

DDP 391 San Juan Puerto Rico 1 KW 8 OCT 2009 eve 400 Learav SSE NNW BFD DX 1766 mi

Here, in Table 346-1 format, are the results of the Log's complete "text" versus its Vocabulary:

 I     | Na |   Ns  |   Sa   |  Ng  | Gr |  Vg  |  ps |   pF  |   Bd |   Dd  |   2R   |   3R   | (2R+3R)/2
 t:1LW | 64 |  2413 |   37.7 |   752| 12 | 3.21 | 02a | 0.228 | 4096 | 0.096 | 0.1202 | 0.0000 |^ 0.0601
 v:1LW | 64 |  1096 |   17.1 |   288| 12 | 3.81 | 023 | 0.206 | 4096 | 0.096 | 0.0693 | 0.0082 |  0.0388

We see that as the Vocabulary morphs into the text, the AvPP2+3FR = {( PP2FR + PP3FR ) / 2} number rises,
hence the metaphorical "anti-text" peculiarity of the Voynich Manuscript samples, within the confines of the data
of Table 346-1, continues.

Berj / KI3U

[1] vms-list post by Berj / KI3U : Thu 7/29/10 10:00 PM
RE: VMs: Version V1.0 of the JVSCGS Chaotic Groups Generator and Utilities computer program

[2] Journal of Voynich Studies communication #346 (Vol. IV, 21 JUL 2010) by Berj / KI3U :
J.VS: Initial experimental results: n-graph spectroscopy of texts and Vocabularies

[3] Journal of Voynich Studies communication #347 (Vol. IV, 26 JUL 2010) by Berj / KI3U :
J.VS: The JVSCGS Chaotic Groups Generator and Utilities computer program

From: Berj N. Ensanian KI3U
To: Journal of Voynich Studies
Sent: 08-7-2010 11:27:07 PM

J.VS: A closer look into the "anti-text" peculiarity of the Voynich text

Dear Colleagues

In J.VS communications #346 [1], in my comments on the experimental results presented there, I remarked
on the peculiar contrast between the di- and tri-graph characteristics of the Voynich text in both small and
large blocks, compared with the non-Voynich text samples, also both small and large blocks :

" ..... So then, within the confines of the data in the table, the Voynich Manuscript writing behaves like
a kind of "anti-text". ..... on the folios it gives the impression of some sort of writing, albeit in
unfamiliar script, yet it in some sense it is an anti-text, this anti-text characteristic being even more
pronounced in the last part of the book compared with the early part. "

Later in J.VS comm. #348 I added an analytic datum to the relevant Table 346-1 of [1], with the n-graph
spectroscopy summary of a technical Log [2].

Whether or not this antitext property continues to hold up as more text samples are analyzed, it is already
of interest to understand just how this peculiarity comes about - its actual text mechanics. Here I would
like to start on that. As we have been, throughout this n-graph spectroscopy work, when handling Voynich
text we use the transcriptions and transcription system of well known long-time Voynich researcher
Glen Claston (GC).

In all the following I have double-checked the numbers, so hopefully there are no errors. The material is
divided into these sections:



Our specific focus is the portion of Table 346-1 under the heading :


Let us first recap how the observed anti-text peculiarity arises, out of the fairly simpe n-graph
spectroscopy analysis track employed in 346-1-E :

1.) The Vocabulary of a block of text is extracted. [3]

2.) The AvPP2+3FR, being the average of the prime permutation digraphs frequencies ratio PP2FR, and the
prime permutation trigraphs frequencies ratio PP3FR, is computed for the text [4], and again for its
Vocabulary. Recall that the PP2FR and PP3FR of a text or Vocabulary are computed upon their respective
prime-symbols, those being the two, and three most-frequent symbols.

3.) These two numbers, t-AvPP2+3FR of the text, and v-AvPP2+3FR of the text's Vocabulary, are compared.
The comparison is imagined as "the Vocabulary morphing into the text". Alternatively, we could have the
analysis go the other way, and the text morphs back into its genetive Vocabulary.

4.) To the extent of the available data so far, we find that :

4a.) with chaotically-generated pseudo-random "texts" the t-AvPP2+3FR and v-AvPP2+3FR numbers are the same.
Hence the chaotic pseudo-random texts serve as a reference.

4b.) with ordinary texts, Askam's English, Vulgate Latin, American-English stammering, Evelyn's English,
and a highly-abbreviated technical Log, the t-AvPP2+3FR is greater than the v-AvPP2+3FR.

4c.) with various Voynich text blocks, large and small, the t-AvPP2+3FR is less than the v-AvPP2+3FR. Hence
the metaphorical denoting of VMS text as "anti-text" - it looks like good text, but morphs oppositely.

Why? Why, with Voynich, does the AvPP2+3FR decrease as the Vocabulary morphs into text? What causes it,
and is the cause a possible clue to the genesis of the Voynich Vocabulary?

Even before we attack this problem in detail, our familiarity with the AvPP2+3FR numbers as special summary
characteristics of the n-graphs spectra of the text and its Vocabulary, has us suspect that there operates
a rule during the morphing process which prevents certain words in the Vocabulary from appearing often in
the text, while favoring other words for duplication in the final text. Whether or not that rule is deemable
as "grammar" is, at this stage, a separate matter.

It is clear though, that our attack must seek to find out what is, from the perspective of AvPP2+3FR,
fundamentally different between rare or unique words, and repeated words, in the final text.


Let us attack the problem in three major steps as follows:

II-1.) We isolate the groups / words in the VMS text samples which are responsible for driving down the
AvPP2+3FR during morphing of Vocabulary to text.

We will do this by subtracting the Vocabulary from the text : the subtraction is carried out as 1 to 1 groups only,
so that Vocabulary words which appear in the text n times, appear in the remainder n-1 times.

To clarify this subtraction process, consider the fourth paragraph above,
"Whether or not this ....... to start on that. " 
and take it as a block of text. Extract its Vocabulary, and do the 1 to 1 subtraction:

Table 349-1

text-list {t} : Vocabulary-list {V} : remainder of subtraction {z} = {t} - {V}

Whether    : Whether    :
or         : or         :
not        : not        :
this this  : this       : this
antitext   : antitext   :
property   : property   :
continues  : continues  :
to to to   : to         : to to
hold       : hold       :
up         : up         :
as         : as         :
more       : more       :
text text  : text       : text
samples    : samples    :
are        : are        :
analyzed   : analyzed   :
it         : it         :
is         : is         :
already    : already    :
of         : of         :
interest   : interest   :
understand : understand :
just       : just       :
how        : how        :
pecularity : pecularity :
comes      : comes      :
about      : about      :
its        : its        :
actual     : actual     :
mechanics  : mechanics  :
Here       : Here       :
I          : I          :
would      : would      :
like       : like       :
start      : start      :
on         : on         :
that       : that       :

As we can see, the list set {z} contains just four words remaining after the subtraction process.

II-2.) On the sets {z} we compute for inspection some n-graph spectroscopy numbers; we will be most
especially interested in the permutation digraph and trigraph frequencies ratios, these computed upon not
their own, but on the tri-prime symbols of the Vocabulary, and tri-prime symbols of the text.

Lets clarify this: above in {z} the di-prime symbols set is comprised of t and o in rank order. But the
tri-prime symbols set is actually, as it expectedly happens in this particular very small {z} case, severely
n-furcated, and we have 5 equally valid ones:


But this complication doesn't matter because we are presently not interested in the prime-symbols of the {z} set.
Instead we are interested in the prime-symbols of the original text and Vocabulary - it is after all
their AvPP2+3FR numbers we are investigating.

Hence we will, for {z}, compute not the prime-permutation, but rather just permutation digraph and trigraph
frequency ratios, using the prime-symbols of {t} and {V}. We will for {z} compute P2FR and P3FR numbers,
rather than its PP2FR and PP3FR numbers.

(Our present nomenclature is still developmental and no doubt can be improved. We might for example write
"teP2FR" and "teaP3FR" and it would be clear enough what we mean, but the variable-names are then getting
longer. We'll do the best we can as we go along.)

For the sake of completing this clarifying example, the tri-prime set for {t} is tea, and for {V} it happens
to be also tea. Therefore we would for example, be computing the permutation digraph frequencies ratio,
P2FR, for {z}, upon t and e, and not on t and o. Similarly with the P3FR, computing it upon t and e and a,
rather than upon any or all of the five spectral n-furcations above.

II-3.) Assemble into a comparisons table the data obtained in II-2.) along with some data already in Table 346-1.
At first it will be guesswork as to which items to include in this new table, and which items to leave out -
we want to zero-in on the mechanism responsible for the anti-text effect without un-necessary clutter.

To summarize steps II-1.) - II-3.) : for the set {z} of those Vocabulary words which are responsible for the
anti-text effect, compute for inspection the permutation digraph and trigraph frequencies ratios based upon
the prime symbols of the Vocabulary, and again based upon the prime symbols of the text that the Vocabulary
morphs into.

Proceeding so, here follows the data obtained with steps II-1.) - II-3.). For the VMS text blocks we've
taken the exact same ones which are in Table 346-1-E. We've included in addition to the VMS samples also one
non-VMS sample from 346-1-E, similarly processed as a check on the II-1.) - II-3.) procedure : it is the
Askam Menta Rubea sample, that one being the closest of the non-VMS cases to the VMS ones, insofar as goes,
so-to-speak, the "distance" between text and anti-text characteristic, per the measures of 4b.) and 4c.) above.

Table 349-2 [5]

change:  ^ = upward   _ = ~none   v = downward

   I        |   157   |   25R   |  v58R   |   79V   |   316   ||    M    |
 tNa        |   149   |    26   |    50   |    41   |   100   ||    34   |
 vNa        |   149   |    26   |    50   |    41   |   100   ||    34   |
 zNa        |    61   |     7   |    17   |    18   |    52   ||    25   |
 tNs        |  37938  |   180   |   1651  |   1534  |  46287  ||   2403  |
 vNs        |  16258  |   161   |   1333  |    972  |  17078  ||   1220  |
 zNs        |  21680  |    19   |    318  |    562  |  29209  ||   1183  |
 tNg        |  10261  |    49   |    399  |    383  |  11816  ||    594  |
 vNg        |   3518  |    42   |    292  |    226  |   3572  ||    235  |
 zNg        |   6743  |     7   |    107  |    157  |   8244  ||    359  |
 ztRNg (ATR)|  0.657  |  0.143  |  0.268  |  0.410  |  0.698  ||  0.604  |
 zvRNg (AGR)|  1.917  |  0.167  |  0.366  |  0.695  |  2.308  ||  1.528  |
 tVg        |   3.70  |   3.67  |   4.14  |   4.01  |   3.92  ||   4.05  |
 vVg        |   4.62  |   3.83  |   4.57  |   4.30  |   4.78  ||   5.19  |
 zVg        |v  3.22  |v  2.71  |v  2.97  |v  3.58  |v  3.54  ||v  3.30  |
 ztRVg      |  0.870  |  0.738  |  0.717  |  0.893  |  0.903  ||  0.815  |
 zvRVg      |v 0.697  |v 0.708  |v 0.650  |v 0.833  |v 0.741  ||v 0.636  |
 Tps        |   o91   |   oa1   |   aoe   |   o9c   |   oa9   ||   etn   |
 Vps        |   o91   |   o19   |   oae   |   9oc   |   oc9   ||   ets   |
 Zps        |   o91   | a8m/am8 |   aeo   |   o9h   |   oa9   ||   ten   |
 tP2FR      |  0.0019 |  0.0000 |  0.0000 |  0.0000 |  0.0014 ||  0.0592 |
 tP3FR      |  0.0042 |  0.0000 |  0.0094 |  0.0000 |  0.0000 ||  0.0659 |
 tAvP2+3FR  |v 0.0030 |_ 0.0000 |v 0.0047 |v 0.0000 |v 0.0007 ||^ 0.0626 |
 tAvPP2+3FR |  0.0048 |  0.0000 |  0.0252 |  0.0007 |  0.0016 ||  0.0573 |
 vP2FR      |  0.0019 |  0.0000 |  0.0000 |  0.0000 |  0.0528 ||  0.0592 |
 vP3FR      |  0.0042 |  0.0000 |  0.0094 |  0.0000 |  0.0000 ||  0.0583 |
 vAvP2+3FR  |v 0.0030 |v 0.0000 |v 0.0047 |v 0.0000 |v 0.0264 ||^ 0.0588 |
 vAvPP2+3FR |  0.0070 |  0.0373 |  0.0300 |  0.0011 |  0.0415 ||  0.0460 |
   I        |   157   |   25R   |  v58R   |   79V   |   316   ||    M    |


The data, as seen in the bottom eight rows, confirms our above initial suspicion (due to our familiarity
with the AvPP2+3FR numbers) about the origin of the anti-text characteristic:

The VMS Vocabulary morphs into text bound by a rule which suppresses the proliferation of Vocabulary words
which would contribute to an increase in AvPP2+3FR. As seen in the table, the mechanism works across a great
range of text gross-characteristics - in particular it is interesting that only in the case of the large 157 block are
the Tps, Vps, and Zps completely identical.

We might object to denoting the condition here with the term "suppresses" - we have not proved that the
observed condition, at this point based on limited data, could not be simply accidental. But we do have
enough data in Tables 346-1 and 349-2, to justify using this term in its neutral or mathematical sense.

So that then is the underlying mechanism relating the Voynich text with its Vocabulary, for the cases
studied, which together account for a considerable portion, about two-thirds, of the total Voynich text. Let us
better comprehend the meaning of this data in Table 349-2 by way of analogy, where we imagine that we are
teaching a class on essay composition:

We hand our class's students a Vocabulary list which has some proportion of the words marked with an
asterisk. We tell the students that they are to write an essay, of whatever length they like, short or long,
with the following rules applying:

a.) the essay must use only words from the provided Vocabulary list
b.) every word in the Vocabulary must appear in the essay at least once
c.) the words marked with an asterisk must appear exactly once in the entire essay

Unbeknownst to our students we determined the asterisk-marked portion of the Vocabulary, not from literary
considerations, but from mathematical considerations: we determined the Vocabulary's tri-prime symbols set,
that is its top three most-frequent alphabet symbols, formed from them the four possible prime permutation
digraphs and the six possible prime permutation trigraphs, and then marked with an asterisk those words in
the Vocabulary which contain these ten n-graphs, thus guaranteeing that as the Vocabulary is morphed into
various essays by our students, their essays will evaluate to AvPP2+3FR numbers which are less than the
AvPP2+3FR of the Vocabulary. [6]

Our analogy is slightly simplified - depending on how a VMS text-block is identified from the entire VMS
text corpus, rule c.) may be slightly relaxed. Alternatively a more complicated rule may be substituted, for
example that for every use in the essay of an asterisk-marked word beyond the first use, three or more un-
marked words of similar word-length must be added into the essay; and this rule may be further refined by
distinguishing the marked words as to their prime digraph and prime trigraph contents, and assigning
specific x2 and x3 multipliers for unmarked words required to co-add, so as to ensure suppressing a rise in

These rules are a reflection of the scheme which created the VMS text. Surprisingly, they ring somewhat
reminiscent of exercises which essay composition professors sometimes assign their students (I say this
based on my own schooldays experiences), except that in the VMS the rules are not, at least on the surface,
literarily-determined rules, but rather mathematically-determined rules.

Why these particular mathematical rules with the VMS text? Do they suggest the presence of a cipher mechanism?


To continue our probing of this anti-text mechanism let us seeing if we can find some additional noticable
patterns among the Voynich groups which function as AvPP2+3FR suppressors and anti-text makers, comparing
them with the other Voynich groups. We will collect the {z} sets of the VMS examples in Table 349-2 and have
a look at them. But even before we do the collecting, we are struck by the particular {z} set of 25R (Voynich folio f25r) :

Table 349-4 : n-1 {z} remainders set


It is dominated by none other than the most famous / infamous Voynich text-group and sub-group of all, the
notorious GC-8am (EVA daiin) which Voynich text researchers have been puzzling over for decades. Across the
VMS more than 3% of the words are either 8am or include it. In Voynich f25r this 8am or daiin word is the
chief AvPP2+3FR suppressor, the chief anti-text maker.

One wonders, as has been wondered in the past, if this 8am group is a null-group filler. Indeed it is easy
to raise the question: are all AvPP2+3FR suppressors just fillers, to be deleted from a block of VMS text,
in order to recover the real payload portion of the text? [7]

It is curious too that in Table 349-4 the group GC-Kan is a member of {z} - this connects with long puzzled-over
questions about the nature of the complex compound or intruding gallows glyphs in the VMS text, the ones which
contribute so greatly to the Voynich text's beauty; again as has been wondered in the past, are they just decorations?
We do pause to remind ourselves that the transcription system in use is critical to our proceedings.

Needless to say, one could turn the perspective around and ask: are the AvPP2+3FR suppressors the payload
groups? But considering how common 8am is in the text, especially in the front botanical portion of the VMS,
this track would invite considering a cipher scheme to be based not just on the specific compositions of
groups, but also the locations they mark in the text-grid, and the context of the other groups around them.
A variation is that anti-text groups serve as keys of some sort to the normal groups, or conceivably vice versa.

Now to probing further into the nature of the anti-text maker groups - this is likely to involve some
extensive effort (considerably eased with [4]), and we'll just make a start on it here.

Total number of groups in the VMS text samples = 22908
Total number of groups in the VMS text samples' Vocabulary = 6352 [8]

this makes for a Vocabulary Ratio in the samples = 6352/22908 = 0.2773

Total number of {z} anti-text groups = 15258

Text-level Anti-text Ratio for combined VMS samples = ATR = 15258/22908 = 0.6661

Impressive! Clearly the bulk of the VMS text, across our samples, is composed of groups suppressing the 
AvPP2+3FR. The table of course indicates that the ATR varies greatly across specific text blocks, so that in
f25r the ATR is only 0.1429, but in 316 (f103R-f116R) the ATR = 0.6977 . It's inviting to consider the ATR
number as a means of classifying Voynich text blocks.

The total of anti-text groups of course contains many duplicates. Lets obtain the Vocabulary specific to the total {z}'s :

Size of the anti-text Vocabulary = 1602 groups

So, these 1602 basic anti-text makers are inflated by a factor of almost 10 across our text samples, as they do their
work of suppressing the AvPP2+3FR. These 1602 groups range in size 1-7, the most common right in the middle being
a size of 4 glyphs. The 1602 groups take up 69 VMS alphabet glyphs, with the 17 most-frequent, in rank order, being:

GC-o, 9, c, 1, a, 8, h, e, k, y, 4, C, 2, m, 7, s, g.

This set is very similar to other 17-most-common-glyphs sets (specific VMS-block dependent). The most common
digraph across the 1602 groups is GC-89, and the 1602 includes its flagship group: GC-898989 . There are
plenty of those well-familiar groups which differ by just one glyph, for example:

Table 349-5

GC-hc89, kc89, 1c89, 2c89, 3c89, 5c89, Gc89, Hc89, Kc89, %c89

GC-okae, okam, okan, okap, okay, okax, okaz, okaN, oka*

GC-4okae, 4okam, 4okan, 4okap, 4okay, 4okaz, 4okaM

But the normal text groups also have plenty of these differ-by-1-glyph examples. We will see some light shine on that shortly.

Lets us now compare the Vocabulary of the combined samples with its sub-Vocabulary, the anti-text groups Vocabulary.
We find that they both have the same tri-prime symbols set and in the same rank order: o9c

We further see:

Vocabulary-level Anti-text groups Ratio = AGR = 1602/6352 = 0.2522

So a quarter of the Vocabulary is comprised of anti-text maker groups, which as we have seen, are as a whole
inflated by a duplication factor of nearly 10 in the final collection of the VMS sub-texts, far more inflation than with
the normal groups : (22908 - 15258)/6352 = 1.2044 . But Table 349-2 shows that just as with the ATR, the AGR
varies widely across specific blocks.

Now let us separate the Vocabulary of our collected samples into its two sub-Vocabularies: the normal text
sub-Vocabulary, and the anti-text maker sub-Vocabulary - they will have no groups in common. Lets compare them
according to the format of Table 346-1 :

Table 349-6

 I     | Na |   Ns  |   Sa   |  Ng  | Gr |  Vg  |  ps |   pF  |   Bd |   Dd  |   2R   |   3R   | (2R+3R)/2
 Norml | 153| 23942 |  156.5 | 4750 | 10 | 5.04 | o9c | 0.330 | 23409| 0.045 | 0.0074 | 0.0013 |  0.0044
 Anti  |  69|  6682 |   96.8 | 1602 |  7 | 4.17 | o9c | 0.361 |  4761| 0.063 | 0.0030 | 0.0000 |  0.0015

Table 349-6 confirms everything we know to this point. And, if we wanted to create mathematical operators
which inter-transform normal groups and anti-text groups, then Table 349-6 suggests some of the requirements.
The chief characteristic required of an anti-text maker group is, that when it joins a Vocabulary and text, that it
not drive up the AvPP2+3FR = (2R+3R)/2 , and whenever possible to lower it for the Vocabulary and text.
In that, keeping the PP2FR = 2R down is more difficult than keeping down the PP3FR = 3R because as every glyph,
especially a prime glyph, is added to a proto-group and increases its n-graphs content, it is far more likely that
a prime permutation digraph will arise, than a prime permutation trigraph. Hence, one obvious mechanism for generating
anti-text maker groups is to keep them short - we see in Table 349-6 that the anti's are almost one glyph shorter on average
than the normals: 4.17 versus 5.04 .


Remove from the normal group a non-prime symbol (so as to keep the primality of the prime symbols intact) in such a manner
so that in the resulting group's new n-graphs spectrum there do not arise more prime permutation digraphs and trigraphs.

Lets try it - from the Vocabulary of the normals we take the group GC-4okaie and remove GC-i to obtain:


which is indeed an anti-text maker as we have seen above in Table 349-5. And, our tentative rule has been
adhered to: the GC-i is not a prime symbol (the primes are o9c), and on removing it the one new created
digraph GC-ae is not a prime permutation digraph (those being oo, o9, 9o, 99), and also the new created
trigraph GC-kae is not a prime permutation trigraph (those being o9c, oc9, 9co, 9oc, co9, c9o).

Looking over the normal and anti- Vocabularies (of our collected samples, which although a large fraction,
are not the total VMS inventory) we will see many of these kinds of related groups, for example:

Table 349-7

normal : anti

18c89 : 1c89
71c89 : 1c89
1ec89 : 1c89
j1c89 : 1c89
1kc89 : 1c89
s1c89 : 1c89
u1c89 : 1c89
f2c89 : 2c89
h2c89 : 2c89
k2c89 : 2c89
s2c89 : 2c89
e3c89 : 3c89
k3c89 : 3c89
e5c89 : 5c89
85c89 : 5c89
4Hc89 : Hc89
Ahc89 : hc89
e%c89 : %c89

It is, by analogy, as if we are dealing with particles and anti-particles, matter and anti-matter. The tentative rule for
the inverse transformation is suggested:


Add to the anti-text group any symbols which either conserve or increase the prime digraphs and prime trigraphs content in the result.

Lets try it, while staying within our data sets :

Table 349-8

 anti : normal

29    : 2co9
59    : 5co9
k9    : kco9
c9    : occ9
1y    : 1c9oy
1oy   : 1c9oy
hc9   : hcco9
hc9   : ochc9
ohC9  : gocohC9
ohoy  : ohocoy
4oh9  : 4ohco9
o8am  : 9o8am
o8am  : 2co8am
o8am  : 3co8am
1ch9  : 1cooh9
1oH9  : oo1oH9
4okae : 4okaie
4okam : o4okam

With just paper and pencil handwork, that is working conditions of centuries ago, it appears to be easier
and faster to create a starting Vocabulary of normals, find out its tri-prime symbols set, and then
transform some portion of this normal Vocabulary into an anti-Vocabulary set. Essentially this would the
exercise of the essay composition professor analogy.

Obviously these transformation rules permit generating many more rule-valid groups than the finite Voynich
text corpus contains. Also conversely, we have not proved that these rules permit inter-transformation of
all known Voynich groups. For example, we can contemplate complicating modifications of the rule which
transforms normals into anti's, by specifying additional rules for removing also a prime symbol from the
normal group.

And also, even though this far in investigating the anti-text effect we have been lucky with our chosen VMS
text samples, individually, and collected, the old problem of proper identification of a VMS text-block for
purposes of analysis, earlier mentioned in [3], will remain, and it relates directly to the question: are the tri-prime
symbols which are the key for distinguishing normal and anti-text groups local and therefore variable, or are they
global and fixed for the entire text corpus? It would require very extensive experimentation with many different
text blocks to obtain a firm idea on that.


And this brings us back to the question: what is the anti-text effect a reflection of, in the composition of
the Voynich Vocabulary and text? If the effect is not accidental, but a direct reflection of a VMS
Vocabulary and text generating algorithm, then what clues does it give about the algorithm?

The first obvious clue of course is that the VMS groups fall into two distinct classes - normal and anti-text.
And as already pondered above, from this the simplest possible suggestion is that one class is decoys
or filler-nulls, and the other class holds the payload text. If as mentioned a moment ago the sheme began
with generating the normal Vocabulary, then we could have two concealment techniques operating side-by-side
in the VMS:

the "normal" Vocabulary is the result of group-level encipherment of plaintext words, and the anti-Vocabulary
supplies steganographic decoys.

Proceeding along the simplest possible track, the group-level "encipherment" of the normals would be just
mapping them to a strange alphabet, in this case the VMS alphabet. If so, then plaintext recovery would be a
matter of matching the readily available letters-frequencies statistics of the normals in a block of text,
to those of known samples. Alternatively the method of group frequencies by Jan Hurych could be tried [9].

Lets just have a quick look at the short botanical f25r text from the perspective that its anti-text groups
are filler-nulls to be thrown out. We'll take its transcription from GC's voyn_101.txt and strike out all
the anti-text groups ( 8am and so on from Table 349-4 ) to obtain for its six lines:

Table 349-9

 f1oe89 so29 Eh9 3o89 o1oe89 Goe89.s9
 okoy 1oy 1sh9 1ok19 2az 4o8 2o 19 h19 1han
 4ok19 4ok29 1CsCs 2Cay 8cn
 81H9 2oK9 9k1c9 Koy 1aN 1am 4ok1an
 4ok1cam 81an 4okam
 ohae 1okam 8aiy ok aZ okos9

It's altogether very little data, but lets just see what the stats look like:

Table 349-10

 No. of Groups =  38
 Total no. of symbols =  151
 Average Group-size =  3.973684
 Total-Symbols-to-Alphabet Ratio =  5.8077

 Frequencies of the groups :

 group-size : frequency

 02 :  5
 03 :  9
 04 :  11
 05 :  9
 06 :  3
 07 :  1

 Frequencies of the symbols :

 Rank : Symbol : Frequency : Frequency-percentage

 01 : o :  24  :  15.89404
 02 : 1 :  19  :  12.58278
 03 : 9 :  18  :  11.92053
 04 : a :  13  :  8.609271
 05 : k :  11  :  7.284768
 06 : 8 :  9   :  5.960265
 07 : 2 :  6   :  3.97351
 08 : 4 :  6   :  3.97351
 09 : s :  6   :  3.97351
 10 : h :  5   :  3.311258
 11 : y :  5   :  3.311258
 12 : e :  4   :  2.649007
 13 : m :  4   :  2.649007
 14 : n :  4   :  2.649007
 15 : C :  3   :  1.986755
 16 : c :  3   :  1.986755
 17 : K :  2   :  1.324503
 18 : 3 :  1   :  0.6622517
 19 : E :  1   :  0.6622517
 20 : G :  1   :  0.6622517
 21 : H :  1   :  0.6622517
 22 : N :  1   :  0.6622517
 23 : Z :  1   :  0.6622517
 24 : f :  1   :  0.6622517
 25 : i :  1   :  0.6622517
 26 : z :  1   :  0.6622517

It's interesting that there are 26 alphabet symbols. From long experience with simple direct mappings of
letters we already know what to expect - a mix of actual, and maybe-words, which can be worked further with
separations, concatenations, de-anagramming, de-abbreviating, re-arranging and so on, but just out of
curiosity to see the effect of the removal of the anti-text groups, lets try a simple mapping, one which
double-maps many of the letters and brings the final alphabet count into the neighborhood of 17 :

Table 349-11

 Rank : f25r GC- : Latin Letter

 01 : o :  I,J
 10 : h :  I,J
 02 : 1 :  Q
 03 : 9 :  A
 22 : N :  A
 04 : a :  U,V
 09 : s :  U,V
 16 : c :  U,V
 21 : H :  U,V
 05 : k :  N
 06 : 8 :  E
 14 : n :  E
 07 : 2 :  S
 08 : 4 :  C
 11 : y :  O
 15 : C :  O
 12 : e :  R
 25 : i :  R
 13 : m :  M
 17 : K :  M
 18 : 3 :  G
 19 : E :  D,M
 20 : G :  L
 26 : z :  B
 23 : Z :  T
 24 : f :  T',D'







As we see, as is often the case with these simple exercises in the Voynich, the result can give the
impression of poetic rhythm, which then recalls Captain Currier's assertion that the Voynich text-line is a
functional entity. Studies which focus on group-counts versus lines, would note that 7 groups each in the
1st and 4th lines, precede a pair of lines that are in 2:1 or 1:2 relationships. And then there are additional numbers
curiosities of the kind which turn up so abundantly in the VMS - lets tabulate the groups counts versus line number
of Table 349-11 to see some of them:

Table 349-12 : f25r normal groups

line-no. : group-count

1 : 7

2 : 10
3 : 5

4 : 7

5 : 3
6 : 6

10+3=13; 5+6=11; 10-5=5; 6-3=3; 7+10=17; 7+6=13; (7-1)+(10-2)+(5-3)+(7-4)+(3-5)+(6-6)= 17

and so on and so forth, producing prime numbers by the cart-load, as the counting of objects in the VMS so
routinely does. (Prime numbers not to be confused with prime symbols.)

Lets have a quick look at the anti-text groups of f25r :

Table 349-12 : f25r anti-text maker groups

line-no. : group-count : groups GC-

1 : 2 : 8am 8am
2 : 0 :
3 : 4 : s 1an 8am 1an
4 : 1 : s
5 : 4 : Kan 8am 8am Kan
6 : 0 :

This table seems to prefer 0, 1, 2, and 4, but it hosts a total glyphs count of 29, which is a prime number.
And that is interesting, because the total glyphs count of the normals is 151, also a prime number.

So, the dividing of the total VMS f25r text into its normal and anti-text sub-texts, proceeded such that the
total glyphs count, 180, divided into the two primes 29 and 151. Once more, we remind that all this is
highly dependent upon the transcription SYSTEM, in this case the system created by GC. And the merits of
various Voynich transcription systems have been extensively debated over the years, and occasionally still
continue to be debated.

From my point of view, the anti-text peculiarity of the Voynich text is one more in a growing series of
indicators that the VMS creator was, aside all else, quite a mathematician, someone who really liked numbers
and their relationships. And that the VMS author created some sort of math-machine for generating his
Vocabulary and text, and that even if its purpose was to encipher plaintext, that at least as much his
algorithm was so constructed to display his love of numbers and their relations, including their relations to text.
And that's different from just devising a clever system for hiding plaintext - it's different because the algorithm projects
a philosophy, and therefore it projects a philosopher behind it, not just a sharp technician.

Berj / KI3U

[1] J. Voynich Studies communication #346 (Vol. IV, 21 JUL 2010):
J.VS: Initial experimental results: n-graph spectroscopy of texts and Vocabularies, by Berj / KI3U.

[2] J. Voynich Studies communication #348 (Vol. IV, 30 JUL 2010):
J.VS: n-graph spectroscopy summary of a Log, by Berj / KI3U.

[3] In the case of the Voynich Manuscript determining a text-block isn't as straighforward as it sounds
because it is technically uncertain what constitute the boundaries for a proper block of text suitable as an
analytic unit. For example, commonly we take the "paragraph" to be a text-block which on its page appears to
have some free space setting it apart, and has its first word on its first line begin with a gallows letter,
or sometimes begin with the digraph GC-4o. But this choice breaks down here and there in the folios, and
even early on in the book, on f10r we see a clear example. For another example, on f84r, the bottom
"paragraph" visually suggests it is a paragraph, and indeed GC in his voyn_101.txt transcribes it as such,
however it starts not with a gallows letter, but with an ordinary GC-2. So is this really its own paragraph,
or just a continuation of the paragraph above it? In our work here it obviously makes sense to take as the
text-blocks the exact same ones which first indicated the anti-text effect under investigation.

[4] Journal of Voynich Studies communication #347 (Vol. IV, 26 JUL 2010) by Berj / KI3U :
J.VS: The JVSCGS Chaotic Groups Generator and Utilities computer program
The current version V1.0 will soon be replaced with updated version V1.1 .

[5] Definitions and Identifications pertaining to Table 349-2 (see also [1] for additional details) :

Source Identifiers I :

157 : VMS f1r-57r from GC's voyn_101.txt, soft-spaces made hard
25R : VMS f25r from GC's voyn_101.txt, soft-spaces made hard; Currier-A
58R : VMS f58r from GC's voyn_101.txt, soft-spaces made hard; Currier-A
79V : VMS f79v from GC's voyn_101.txt, soft-spaces made hard; Currier-B
316 : VMS f103r-116r from GC's voyn_101.txt, soft-spaces made hard
M   : Askam's Menta Rubea


{t} == the set of groups of the text
{v} == the set of groups of the text's Vocabulary

tNa == size of the block/text's Alphabet
vNa == size of the text's Vocabulary's Alphabet
zNa == size of {z}'s Alphabet

tNs == total number of symbols in the text
vNs == total number of symbols in the text's Vocabulary
zNs == total number of symbols in {z}

tNg == number of groups in the text
vNg == number of groups in the text's Vocabulary
zNg == number of groups in {z}

vtRNg == vNg/tNg (ratio of number of groups, Vocabulary to text; i.e. Vocabulary ratio)
ztRNg == zNg/tNg (ratio of number of groups, {z} to text; text-level Anti-text Ratio ATR)
zvRNg == zNg/vNg (ratio of no. of groups, {z} to Vocabulary; Vocabulary-level Anti-text groups Ratio AGR)

tVg == tNs/tNg (text's average group-size)
vVg == vNs/vNg (Vocabulary's average group-size)
zVg == zNs/zNg ({z}'s average group-size)

ztRVg == zVg/tVg (ratio of average group-sizes, {z} to text)
zvRVg == zVg/vVg (ratio of average group-sizes, {z} to Vocabulary)

Tps == tri-prime symbols set in rank order (GC glyphs notation) of the text
Vps == tri-prime symbols set in rank order (GC glyphs notation) of the text's Vocabulary
Zps == tri-prime symbols set in rank order (GC glyphs notation) of {z}

tP2FR == permutation digraphs frequencies ratio of {z} based on Tps
tP3FR == permutation trigraphs frequencies ratio of {z} based on Tps
tAvP2+3FR == ( tP2FR + tP3FR ) / 2
tAvPP2+3FR = from Table 346-1 the AvPP2+3FR of the text

vP2FR == permutation digraphs frequencies ratio of {z} based on Vps
vP3FR == permutation trigraphs frequencies ratio of {z} based on Vps
vAvP2+3FR == ( vP2FR + vP3FR ) / 2
vAvPP2+3FR = from Table 346-1 the AvPP2+3FR of the text's Vocabulary

[6] Let us illustrate a simple example of the anti-text essay assignment analogy. As the imaginary essay-class
professor preparing the special Vocabulary for our students, let us start by taking from above this text paragraph:

" So that then is the underlying mechanism relating the Voynich text with its Vocabulary, for the cases studied,
which together account for a considerable portion of the total Voynich text. Let us better comprehend the meaning
of this data in Table 349-2 by way of analogy, where we imagine that we are teaching a class on essay composition: "

and prepare its Vocabulary, marked according to the anti-text generating rules:

Table 349-3

* text
* together
* Let
* better
* teaching

We see that there are five marked words: text, together, Let, better, teaching. As it turned out in this
case, in the process of identifying these antitext maker words, the prime permutation digraphs were
overwhelmingly more influential than the prime permutation trigraphs.

According to the essay writing rules, the essay must use these five words exactly once. If, for her essay, 
one of our students coincidentally were to write essentially the original paragraph, the marked word "text"
might give her the most trouble - she may use it only once:

" So that then is the underlying mechanism relating the Voynich text with its Vocabulary, for the cases
studied, which together account for a considerable portion of the total Voynich. Let us better comprehend
the meaning of this data in Table 349-2 by way of analogy, where we imagine that we are teaching a class on
essay composition: "

Compared with the original, the second "text" at the end of the first sentence is not there, but the essay
rules are satisfied. Here the impact of the rules, on coincidentally composing a close copy of the original
paragraph, was not severe - the impacted first sentence is only ever-so-slightly awkward in comparison with
the original.

But now imagine what would happen if she wanted to continue the essay beyond the two sentences: the word
"text" would no longer be permissable - it's allotment has been used up. If she plowed on anyway, trying to
edit the expanding essay, which is discussing the Voynich text, then the anti-text effect would become more
and more noticable. With an expanding and presumably continually edited essay, the anti-text would become
more and more noticable no matter what she were writing about - conceivably, from a literary perspective it
might become so awkward as to resemble gibberish, although an n-graph spectroscopy analysis would quickly
reveal the presence of a mathematical system constraining her composition. And, if our assignment to her had
the additional challenge of discovering an essay best matching the given Vocabulary, she might well just
settle on her first composition of just two sentences. Then, approving of her effort, we might ask her to
look over both her composition, and the marked Vocabulary, and ask her to complete "inverting" her anti-text
composition: she might well decide that the only thing in that vein making sense would be to add a second
"text" at the end of her first sentence. (note: in this example the anti-text word "text" is coincidental,
but illustratively convenient).

[7] see pg. 29 of D'Imperio for more on the long considered possibility of nulls.

[8] note: the Vocabularies of the samples have overlaps, so therefore the grand Vocabulary for all the
samples cannot be obtained by simply adding up the entries for the individual Vocabularies in Table 349-2.

[9] Journal of Voynich Studies communication #334 (Vol. IV, 13 FEB 2010) by Jan Hurych :
J.VS: Hurych on Voynich group frequencies: IS THE VM ENCODED? ( and how to solve it in that case)

From: Greg Stachowski
To: Journal of Voynich Studies
Sent: 08-08-2010 6:38:53 PM

J.VS: Wilfrid Voynich and Bronislaw Pilsudski

I have previously (January 2007) mentioned Off-J the contacts of Wilfrid Voynich with the anthropologist
Bronisław Piłsudski [1], famous for his documentation of the Ainu people [2] and also brother of the former
revolutionary, Polish-Bolshevik War commander and Polish Head of State, Marshal Józef Piłsudski [3]. It is
as yet undetermined whether Voynich knew the Piłsudskis through the revolutionary movement, or directly
(they all came from roughly the same area of Poland/Lithuania), or through some other means.

Apart from pointing to yet another one of his contacts with the rich, famous and powerful of the period (for
example Herbert Hoover [4]), these contacts give us additional glimpses into and dates in the otherwise
poorly-documented pre-Manuscript life of WMV. They are also interesting in that they link WMV directly to
Ainu materials and artifacts (including the sale thereof). This, in turn, suggests the wildly speculative
conjecture that the VMS itself might perhaps have been such an artifact, perhaps resulting from some ancient
European-Ainu contact, brought back by Piłsudski and "fenced" by WMV with a false provenance. With no
evidence for it and plenty of circumstantial evidence against it, this conjecture is not, however, put
forward seriously. It must instead remain on the fringes of VMS theory - but it is perhaps no more
outlandish than the Nahuatl (Aztec) or Chinese conjectures discussed at length on the vms-list.

Pilsudski's contacts with Voynich are occasionally listed by the former's biographers, usually under the
Polish form "Wojnicz" or "Woynicz", for example in the "Calendar of events – the biography of Bronisław
Piotr Piłsudski" [5]. Excerpted from that:


* Beginning of June voyage to London to visit a Japanese exhibition. A stay at the friendly house of the
Wojnicz family at St. Peter Sq. 37.

* After some time Piłsudski relocates himself to an inn at Shepherd Bush Green 45, where a large group
of Japanese participants of the exhibition was staying. Soon after that Stefan Żeromski (a renowned Polish
writer) with his family stay at the same address. Thanks to the assistance of the Wojnicz family, Piłsudski
manages to sell some objects from his Ainu collection.

This includes another of Voynich's addresses, for June 1910: 37, St. Peter's Square, (Hammersmith, London W6).

Recently more details regarding the connections between Voynich and Pilsudski have come to my notice. The
following is Note 248 on p. 774 of “The collected works of Bronisław Piłsudski, Volume 3”, edited by Alfred
F. Majewicz, subtitled “Ainu Language and Folklore Materials 2”.[6][7]

From the description: "Volume 3 is devoted exclusively to B. Pilsudski's Ainu-related materials, for their
most part previously unpublished. In addition, it comprises Pilsudski's research reports on his expeditions,
a superb collection of fifty prayers in Ainu as well as texts and melodies recovered from Pilsudski's famous
wax-cylinder recordings of Ainu-folklore of 1902-1903. The bibliographies printed in volume 1 are
extensively enhanced. Abundant illustrative material is included.”

Note 248, p.774:

He must have taken, however, some of the cylinders with him, for from letters of Lydia Woynicz (= Ethel
Lilian Voynich) to Bronisław Piłsudski we learn that she took pains to organize their presentation to
specialists in London (letter dated November 6, 1909), that such presentation in the Royal College of Music
had to be canceled twice because of Ms. Woynicz's illness (letter dated December 27, 1909), that the
presentation was finally scheduled for “Friday, January 28, 1910”, (letter dated January 22, 1910), that two
listening sessions took place in the Royal College of Music and that “reducing to notation” was in progress
(letter dated March 23, 1910), finally, that the music from cylinders was transcribed into musical notation,
but “Prof. Haddon and the people from the Royal College of Music came to the conclusion that it would
perhaps be impossible to sell the cylinders in England” (letter dated October 21, 1910); an apparently later
undated letter from Mr. W[ilfryd] M[ichał] Woynicz implies that there were also many other attempts to
contact various people and institutions to draw their attention to the cylinder records, but all of them
declined to buy them. The letters are preserved in the manuscript department of the Library of the Polish
Academy of Sciences and Letters (cf. This volume, note 470), call number 4646, vol. 4. Ethel Lilian Voynich
(1864-1960) was an English novelist, author of The Gadfly (1897), a now forgotten novel of Italy's 19th
century fight for independence and unity translated into many languages, and her husband W. M. Woynicz was a
noted London antiquarian and political activist who had escaped from his exile in Siberia; both
substantially supported Bronisław Piłsudski, also financially. Three cylinders were identified at the end of
March 2000 in the National Sound Archives of the British Library in London (in the archives of the late
Charles Samuel Myers) by Tatyana Roon as doubtlessly coming from B. Piłsudski's collection (cf. Roon 2001 in
IINBP 5, 154-161, cf. This volume, p.72 ) ; inscriptions on the boxes suggest a hecire jukara and an ihunki
jukara sung by a young woman aged 22-24 named Inuma of the Western Coast of Sakhalin (cylinders with No No
20 and 23, respectively), and a sinocta ~ sake jukara (cylinder No 27) sung by a male named Sambrok of the
Eastern Coast (on Sambrok see CWBP 1, 330).

As can be seen, this minor footnote to the works of Bronisław Piłsudski in turn provides the Voynich
researcher with a wealth of little details about the life of WMV, including the location of some of his (and
ELV's) letters.

- GS

Note: I initially came across the Piłsudski-Voynich connection sometime in 2006 (the date escapes me), and
first mentioned it and the Ainu conjecture Off-J in January, 2007.


[1] see
[2] see
[3] see
[4] J.VS communication #86,
[5] , hosted by Gdansk University, Poland.

[6] Available online through Google Books at:

[7] This edition published by Walter de Gruyter GmbH, Berlin, 2004, ISBN 3-11-017614-9, apparently as part
(“Documentation 15-3”) of the series “Trends in Linguistics” .(see

From: Berj N. Ensanian KI3U
To: Journal of Voynich Studies
Sent: 08-9-2010 10:48:41 PM

J.VS: Some Wilfrid Voynich Book Trade data from 1904

Dear Colleagues

Below is some data [1] from the journal :

Notes and Queries: A Medium of Intercommunication FOR LITERARY MEN, GENERAL READERS, ETC.
Tenth Series, Volume II, July - December, 1904. London, John C. Francis.

This data will supplement similar data in J.VS communications #335, #336, and #350 (all J.VS Vol. IV).

Berj / KI3U

[1] July 23, 1904, p. 80 :

" Mr. Wilfrid M. Voynich issues another of his short catalogues. Mr. Voynich has now such a large stock that
he finds it impossible to give full descriptions in his bibliographical lists, and has decided to issue,
side by side with those lists, ordinary short ones, the present being the ninth, and, like the earlier, full or rarities. "

August 27, 1904, continuation of second advertisement page :

" Mr. W.M. VOYNICH has transferred his stock of Old and Rare Books from 1, Soho Square, W., to ground-floor
Card Index, grouped under Subjects, of all Books in Stock, which enables Specialists to turn up at once all he has to
offer without laborious reference to Authors. "

October 22, 1904, p. 340 :

" Mr. Voynich, of Shaftesbury Avenue, continues his short catalogues, full of rarities, as usual. Among many
of interest we note Sophocles, 1518, .3l. 15s; Xenophon, 1516, 5l. 5s.; Plutarch, 1618, 15s.; Sir Thomas
Herbert's 'Travels,' 1638, 6l. 6s, (the last part relates how "Madoc ap Owen Gwyneth discovered America
above three hundred years before Columbus"; no copy of this is in the British Museum); Nostradamus, 1563,
1l. 1s.; and 'Reformation der bairischen Landrechte,' 1518, printed on vellum, 25 guineas. There is a good
list of English plays. These include the rare first edition of 'The Spightful Sister,' by Bailey, 1667, 2l. 2s.;
John Banks's 'Vertue Betray'd,' 1682, 2l. 10s. (in the dedication is an interesting reference to
Shakespeare); Henry Carey's 'The Honest Yorkshireman,' 1736, 1l. 1s. (acted for one night only at Drury
Lane: "The company after one night's acting was suddenly interdicted, and the House shut up"); Cavendish's
(first Duke of Newcastle) 'The Humorous Lovers,' 1677, 5l. 5s.; Congreve's 'Way of the World,' first
edition, 1700, 2l. 10s. ; and the extremely rare first edition of Otway's 'Alcibiades,' 1675, 5l. 5s. Much
to interest may be found under Italian Literature, Japan and China, Astrology, &c. "

December 17, 1904, continuation of second advertisement page :

MR. W.M. VOYNICH has an immense Stock of these, all full Indexed. He deals principally in Incunabula,
Bindings, SHAKESPERIANA, and English and French Literature up to the Eighteenth Century. "

From: Berj N. Ensanian KI3U
To: Journal of Voynich Studies
Sent: 08-10-2010 2:42:09 PM

J.VS: Ancient "language" mystery of the Picts; WMV 1903 lecture on printed books

Dear Colleagues

A BBC online news article dated 10 AUG 2010 by Victoria Gill :

touches upon some information-theory-entropy issues (though without using the word "entropy") :

" Ancient language mystery deepens "

" A linguistic mystery has arisen surrounding symbol-inscribed stones in Scotland that predate the formation
of the country itself. The stones are believed to have been carved by members of an ancient people known as
the Picts, who thrived in what is now Scotland from the 4th to the 9th Centuries. These symbols, researchers
say, are probably "words" rather than images. But their conclusions have raised criticism from some linguists. "

" Using the symbols, or characters, from the stones, Prof Lee and his colleagues measured this feature of
so-called "character to character uncertainty". They concluded that the Pictish carvings were "symbolic
markings that communicated information" - that these were words rather than pictures. Prof Lee first
published these conclusions in April of this year. But a recent article by French linguist Arnaud Fournet
opened up the mystery once again. Mr Fournet said that, by examining Pictish carvings as if they were
"linear symbols", and by applying the rules of written language to them, the scientists could have produced
biased results. "

On another topic, here's a 1903 lecture by Wilfrid Voynich :


Berj / KI3U

From: Berj N. Ensanian KI3U
To: Journal of Voynich Studies
Sent: 11-14-2010 6:59:17 PM EST

J.VS: A 1917 speculation of Gaelic as the Voynich text

Dear Colleagues

In this work [1] :

An Introduction To The History Of Medicine
With Medical Chronology, Suggestions For Study And Biographic Data

by Fielding H. Garrison, A.B., M.D.
Principal Assistant Librarian, Surgeon General's Office, Washington, D.C.

Second Edition, Revised And Enlarged

Philadelphia and London, W.B. Saunders Co., 1917
Originally Copyright in 1913, reprinted May, 1914.

The author, in:

Ch. VII : The Medieval Period (1096-1438)

under the heading: Cultural and Social Aspects of Medieval Medicine

ends a paragraph on Druidical medicine in Britain, with:

" A strange MS. of Roger Bacon's, written either in Gaelic or cipher, full of remarkably naturalistic
miniature figurations of pregnancy, suggests some kind of esoteric or mesmeric magic, like that of the Hindu fakirs. 1
1 In the possession of Mr. Wilfred M. de Voynich. "

It would be interesting to determine if Garrison already had this in his 1913 first edition, two years
before the OCT-NOV 1915 first public exhibit of Voynich's cipher manuscript at the Art Institute of Chicago.

We recall that D'Imperio, pg. 69-70, under "Alchemical, Medical, and Astrological Symbols.", comments
briefly on considerations of Old Irish (the progenitor of Gaelic etc.), of which there are inscriptions
in Ogham / Ogam. The Ogam alphabet is considered by some to be a cipher script related to German runic
cipher script.

The study of Ogam in connection with the VMS script comes up now and then, one good reason being that Ogham
has a rich scope of diverse possibilities from the perspective of the mysterious Voynich text, including a
kind of vertical boustrophedon. The well-known Ogham Tract in the Book of Ballymote (1390), explaining
Ogham, shows among other items circular diagrams hosting Ogham. Although it has been variously received,
to put it neutrally, Barry Fell's 1976/1989 work "America B.C." is quite handy as a source of various ideas of
possible use in VMS text research, a spectrum of ideas which otherwise are tedious to collect from more
conservative sources. In the chapter "The Druids of New England" Fell writes:

" For the 17-letter Ogam alphabet of Monhegan, Maine, and Ireland, the chances of independent origin in
these two places are less than one in 300,000,000,000,000. "

Elsewhere in his book Harvard Professor Fell pokes a bit of fun at some of his detractors with a
demonstration of "Yale Ogam".

Berj / KI3U

[1] accessible online via google books.

From: Berj N. Ensanian KI3U
To: Journal of Voynich Studies
Sent: 11-17-2010 1:49:02 AM EST

J.VS: More on Wilfrid M. De Voynich and Gaelic-speculator Dr. Fielding H. Garrison

Dear Colleagues

A couple of days ago we had some brief off-J discussions on the subject of J.VS comm. #353 [1] around the
fact that Dr. Garrison had in the 2nd, 1917 edition of one of his books [2] speculated (i.e. he simply says
so without providing any derivation) that the text of Voynich's Roger Bacon cipher manuscript is "written
either in Gaelic or cipher".

Our initial off-J interest was to determine whether or not Garrison had delivered his Gaelic opinion
earlier, in 1913 in the first edition of his book, with our view of the OCT-NOV 1915 first public exhibit of
the Voynich manuscript in Chicago as a time reference of particular interest.

Our colleague Greg Stachowski noted that the first edition of Garrison's book is available in the British
Library, and our colleague Dana Scott then went better and found a copy of it online, which had escaped my
earlier searches:

Examining the 1913 first edition we could not find any mention of Voynich's manuscript, and therefore any
Gaelic opinion on it, although Garrison does mention both Roger and Francis Bacon several times.
We temporarily left the matter there, and I concluded:

" So presumably it was sometime after the 1913 galley-proofs-acceptance of his book that Garrison was
exposed to the VMS, or at least felt confident enough to publish comment about it. The question remains:
was his exposure before, or after the OCT-NOV 1915 Chicago exhibition of the VMS?

As I recall, WMV had been privately exposing the VMS to selected individuals before the Chicago exhibit.
John Manly was one of these guys wasn't he? It would be good to get a better picture of that earliest period
- we might even run into some clue or other pointing to just exactly where WMV got the VMS: Mondragone,
Austrian Castle, Libreria Franceschini or where-ever. "

After leaving it such, it occurred to me to do the obvious and search for a direct connection between
Garrison and Voynich. It turns out that they co-authored a paper in 1917, available online (google books):

Figurations of Skeletal and Visceral Anatomy in the Books of Hours
by Wilfrid M. De Voynich and Fielding H. Garrison, M.D.
Annals of Medical History, Volume I, Fall 1917, Number 3.

Garrison was an associate editor of this journal. We note that here is a clear example of Wilfrid not
discouraging, at the very least, his name being published as "De Voynich". The paper itself, though
interesting, seems to have nothing of immediate VMS interest. However in the same volume is a paper by
David Eugene Smith, LL.D. which is well worth reading for background information in VMS studies:

Medicine and Mathematics in the Sixteenth Century

wherein Smith's discussion of seven major influences tending to turn ancient physicians into competent
mathematicians includes optics, comets and astrolabes, Johann Marcus Marci De Kronland, Roger Bacon, Cardano
and many other names familiar from general VMS studies. ( De Voynich, and De Kronland, hmm, interesting ).

Smith writes:

" ..... medicine was a very natural gateway to mathematics in the early universities. A young scholar was
offered four great possibilities in the mediaeval period, namely, theology, philosophy, law, and medicine.
Of course he might follow out his mental bent without taking any of these paths, but these were the enticing
ones. If his taste was for science in any of its branches, the path of medicine was the natural one to
follow, since, as we have seen, medicine had as auxilliary sciences astronomy (which was mathematics par
excellence in those days) and alchemy, and made not a little use of physics. Scientific training, therefore,
found its path of least resistance through medicine. "

" Before the sixteenth century printing had not been sufficiently developed to make it possible to freely
disseminate thought, while after the close of that century superstition began to give way more rapidly then
ever before to scientific inquiry. "

Perhaps other personalities associated with Annals of Medical History besides Garrison had direct contact
with Wilfrid Voynich and may lead to interesting information. Wilfrid did choose a medical stage to formally
introduce his mysterious manuscript in 1921 - the College of Physicians of Philadelphia.

Berj / KI3U

[1] Journal of Voynich Studies communication #353 (Vol. IV, 14 NOV 2010):
J.VS: A 1917 speculation of Gaelic as the Voynich text, by Berj / KI3U

[2] An Introduction To The History Of Medicine
With Medical Chronology, Suggestions For Study And Biographic Data
by Fielding H. Garrison, A.B., M.D.
Principal Assistant Librarian, Surgeon General's Office, Washington, D.C.
Second Edition, Revised And Enlarged
Philadelphia and London, W.B. Saunders Co., 1917

From: Berj N. Ensanian KI3U
To: Journal of Voynich Studies
Sent: 11-23-2010 12:59:13 AM EST

J.VS: A 1442 French MS Bird-glyph possibility, paragraphs, and Organ Guns

Dear Colleagues

Here are some odds and ends of possible interest in VMS work:

1.) Bird-glyph. A week ago on ebay ended the auction of a medieval French one-sheet manuscript by a seller
("lusarag") in the UK, the item going for U.S. $400 :

The MS is described:

Original 15 th century French manuscript on vellum, dated 1442.
Old French of this period is not easy to read but the name Jeahn de Fonte and Pondra (a place?) are clearly
legible as is the date of 1442 (mil xxxx et quarante deux). The document is also signed on bottom right hand
corner. The condition is good with only a couple of small stains and 2 small holes on the left hand side,
remnants of one of the 2 original wax seals is still present . An extremely rare document worthy of further
research. The document measures aproximately 25.5 cms x 8.5 cms. "

As we can see from the pretty good photos, the main writing begins with a glyph which has more than a casual
resemblance to the basic form of the unusual "Bird-glyph" which starts a paragraph on Voynich f1r and is of
continuing interest to us [1]. As the MS seller points out, the writing is difficult to read, but we ask:
is that initial French symbol similar, in the sense of some sort of punctuation-sign, to our VMS bird-symbol,
or is it just a fancy initial ordinary letter which accidentally looks similar to the VMS f1r example? If the
French MS glyph is not an ordinary letter, can this MS give us any further information on the glyph?

2.) Paragraphs. Investigating the Voynich Manuscript from every conceivable angle we of course include its
style of paragraphing. Some interesting historical comments about the ancient Greek paragraphos and on to
paragraphs of modern times are given by the designer Jon Tan (member of the International Society of
Typographic Designers) in his web log of Mon, 9th Jun 2008 :

" The Paragraph in Web Typography & Design "

Tan writes:

" Paragraphs are punctuation, the punctuation of ideas. ..... Medieval punctuation employed a paragraphus—
also known as a “‘gallows-pole’ or upper-case gamma, or § (later ¶)”— to separate ideas in a running
discourse. White space did not punctuate paragraphs until the 17th century. This was the era of Ben Jonson’s
English Grammar, where he recommended the use of syntactic punctuation. Around that time, the practise of
indenting the first line of a paragraph became part of our standard syntax, along with the use of capital
letters for the start of a sentence, and the use of a space between words. "

Well, as we know, across the Voynich text there are lots and lots of white-space separated paragraphs which
start with one of the Voynich big looped gallows letters. Of particular interest might be the Voynich star-
pages f58r and f112v where we also see indentation.

3.) Multiple Tubes. A couple of months back on vms-list in a discussion of astronomical instruments, in a
list-thread launched by funny bunny, I commented [2] :

" Well of course some of us have long wondered if some of the tubes in the VMS suggest astronomical
instruments, lensed or not lensed. As we know, tubes are most prevalent in the nine-rosettes illustration.
There are so many of them clustered there that it is difficult to not entertain non-astronomical ideas about
them, say organ pipes of some sort. The central rosette looks a bit like a fortress surrounded with gun
tubes or cannon. "

There are online some nice pictures of ancient multiple-barrel "organ" guns, and this SEP-OCT 2010 webpage
in the Ethnographic Arms & Armour Forum of :

shows some good photos of such weapons back to the 15th century. Especially interesting is the 15th c.
German "Totenorgel" (Death Organ) because it is a cluster of five tubes: one of the more prominent tube-
clusters in the VMS nine-rosettes foldout is a cluster of five.

Berj / KI3U

[1] Recently on vms-list there was quite some discussion of the VMS f1r bird-glyph and similars in other
manuscripts; see these vms-list threads:

VMs: fr1birdglyph; launched by funny bunny Sun 9/12/10 11:18 AM.
VMs: the bird glyph is in the Cordoza, too...; launched by H.R. SantaColoma Sun 9/12/10 2:04 PM.
VMs: VMs Vasculum; launched by Dana Scott Tue 9/14/10 4:57 AM.
VMs: Bird Glyph; launched by Dana Scott Sun 9/19/10 1:32 AM.
VMs: Bird Glyph/Spanish; launched by funny bunny Sun 9/19/10 5:44 PM.
VMs: FW: Consulta bibliotecaria - Responder [Consulta #5931300]; launched by Dana Scott Mon 9/27/10 8:40 AM.

[2] RE: VMs: 14th century astronomical instrument?; posted by Berj / KI3U Tue 9/14/10 11:09 PM.

From: Berj N. Ensanian KI3U
To: Journal of Voynich Studies
Sent: 12-06-2010 1:41:54 AM EST

J.VS: A Martinus Santinus datum

Dear Colleagues

Concerning Martinus Santinus (ref. J.VS comm. #341, Vol. IV, 21 APR 2010) here's another datum on him,
dating to 1669, found with google books search:

providing the briefest mention of his rank and philosophical concerns.

Berj / KI3U

From: Greg Stachowski
To: Journal of Voynich Studies
Sent:    Thu 12/09/10 10:07 AM

J.VS: Re: A Martinus Santinus datum

Recent off-J discussions have resulted in an approximate translation of the reference [1]
to Martinus Santinus [2] recently brought to our attention by Berj in J.VS comm. #356.

The reference appears in a text included in the unauthorised 20th volume of the collected works of
Theophilus Raynaud, published in Cracow in 1669 (the 19 authorised volumes were published in Lyons in 1665).

This text is concerned with opinions on 'true' martyrs and martyrdom ("De vera Martyrii notione"),
including which manner of death is suitable for someone to be considered a martyr.
In this particular section part, the author presents lists of European theologians, sorted by country,
who have expressed their view on the suitability of death by disease ("per pestem") to martyrdom.

Santinus is listed as one of the "Theologii Germani pro martyrio per pestem", that is, "German (in the 17th
century sense of the term) theologians for [in favour of] martyrdom by disease".

This gives the context for the translation. Thus:

    Martinus Santinus S[acrae] Theol[ogiae] doctor ac professor
    ordinarius, ait sententiam propositam sibi probari

Martinus Santinus, Doctor of Sacred Theology [3], and professor ordinarius [4]
affirms [5] the proposed notion [6], himself being satisfied [that it is]

    non tantum probabilem, sed etiam ut pene certam,

not [just] with great probability, but even nearly certain,


to be approved [7].

Greg S.


[1] J.VS comm. #356, Vol. IV, 06 DEC 2010
[2] J.VS comm. #341, Vol. IV, 21 APR 2010
[3] An academic degree bestowed to this day by pontifical universities.
[4] Professor ordinarius is a professor with a chair.
[5] in the sense of 'supports'.
[6] i.e. that of allowing martyrdom by disease.
[7] by the Church.

From: Berj N. Ensanian KI3U
To: Journal of Voynich Studies
Sent: 12-23-2010 7:53:09 PM EST

J.VS: Comparison of VMS f68r2 stars-diagram with similar diagram in 1698 ASTROSCOPIUM by Wilhelm Schickard Jr.

Dear Colleagues

The Voynich Manuscript astro-panels f68r1 and f68r2 (part of a triple-panel foldout including the PM-curve
bearing f68r3) are similar in layout: a few lines of text on top, followed by circular illustrations
comprised of labeled stars, anchored at the top and bottom with face-bearing circles apparently symbolizing
the sun and moon, those ringed by text, the f68r1 panel having the sun-symbol at top, and the f68r2 having
the sun at the bottom. Allowing for crudity, one could the get the impression that these diagrams, in addition
to their other themes, suggest views obtained through a telescope or astronomical sighting tube.

Recent off-J discussions concerning the interpretation of the f68r3 PM curve with our colleague Robert Teague
motivated me to see what is now available online via google books that might add to our stock of data
on the f68r triple panel. I came across this 1698 work, the scanned copy of which resides in the Bayerische
Staatsbibliothek in Muenchen :

Pro facillima Stellarum cognitione excogitatum, & Commentariolo illustratum
Wilhelmo Schickardo Juniore.
Stutgardiae & Lipsiae,
apud Joh. Herebord Klosium.

Schickard has come up before in past Voynich considerations [1].

The first page, the frontispiece, of Schickard's Astroscopium book is an astronomy-mythology-theme
illustration: in the foreground the Greek mythological Titan Atlas is holding up the circular heavens, while
in the background a philosopher is observing the stars in the heavens with a hand-held telescope.
The circular heavens portion of this illustration, even at first glance, is seen to have great similarity with
the Voynich f68r2 circular astro-diagram, although they are oriented by a rotation difference of
approximately 90 degrees with respect to each other.

Here then is a chance to delve further into the style of the VMS f68r2 (and f68r1) illustration: is there a
tradition? Indeed, our colleague Greg Stachowski remarked to me about this in an email, pointing out that
this is very interesting, in that this is the first time we've seen an image which is similar to those
folios somewhere else, and it is worth noting. Further, Greg wondered how common and at what times the
imagery with both Sun and Moon at either sides of the sphere was.

I undertook a first effort to compare the Schickard and f68r2 diagrams, to see to what extent the basic
patterns made by their respective groups of stars match or approximately match. The results of the first
effort comparison are, I think, intriguing and encouraging.

To effect the comparison I removed the star labels from the VMS f68r2 diagram, and rotated it 90 degrees
c.c.w. I also rotated c.c.w. a few degrees the ASTROSCOPIUM diagram : the respective moon and sun symbols
are now aligned. Using nine different colors I indicated on these the matching patterns. Finally I combined
the two into one side-by-side comparison image:


I have sent to our Librarian (Greg) the .jpg of this comparison image for installation as
Library deposit # 31-1-2010-12-23

As said above, this is just a first effort. Greg noted that some of it looks vaguely like constellations,
but it doesn't really match very well. He further thought that he believes most, if not all, depictions of
constellations before the 17th C were reversed, as on a celestial globe, but this needs to be checked. I had
tried various mirror imagings of the two illustrations during the comparisons. Greg noted that if these are
to be visualisations of a telescopic view, then of couse the immediate problem of Sun, Moon and stars in one
field comes up.

But it is clear that there is rich potential here for gathering more insight into the Voynich's f68r astro-
panels, especially f68r2, with comprehensive exploration of the Schickard family, the ASTROSCOPIUM
(apparently a kind of astronomical clock / tellurium instrument [2]),  and whatever other works by the
Schickards, and other relevant works they lead to, can be found. Schickard Jr. and Sr. look to be our kind
of guys, persons polymath definitely of interest in the pursuit of the greater Voynich mystery. Schickard
Sr. (1592-1635), taught biblical languages, was a working colleague of Kepler, and ahead of his time in
matters map-making and calculating machines [3,4].

As for Schickard Jr., the author of the present book, he seems to have been the nephew rather than the son
of Schickard Sr. [5]

It would seem that a next good step is to get a precise idea of what Schickard means to convey with his
ASTROSCOPIUM heavens diagram, in particular to identify which specific stars and asterisms he is depicting,
if indeed he is, and their various names across the ages (e.g. Schickard used the name "Altor" for Taurus [1]). [6].

Also we might see about determining what may be the name of a previous owner of the Bayerische
Staatsbibliothek copy, handwritten on the book's title page: possibly a Georg Abbas.

Berj / KI3U

[1] for example in "Arabic Star Names", a list from Allen collected by VMS researcher Don Latham :

[2] In Voynich matters we are already well familiar with the great astronomical clock in Prague.


[4] A curious reference to Schickard's working with an apparently somewhat problematic "manuscript of
Vienna" is found in "Miscellaneous Works of Mr. John Greaves, Professor of Astronomy in the University of
Oxford", Volume 1, London, J. Hughs, 1737. It seems it had been left to Schickard to produce a European
distillation of the celebrated Tables of Abulfeda, and that he did so admirably in 1628. However, on page
xxxvi we read this: "But Schickard, in a Letter to Mr. Greaves, observ'd that this manuscript of Vienna,
which he had made use of, was in a variety of places impossible to be read, and generally very doubtful in
the numbers; so that no tables, or at least only very incorrect ones, could be form'd by it."

Greg notes that Schickard Jr. says as much in the text ('patruus' means uncle, father's side).

[6] see also:

6a.) all VMS f68r work by Robert Teague in the J.VS Library and communications archive.


6c.) Richard SantaColoma's 5 DEC 2010 weblog commentary "Is that you, Martin Luther?", wherein Rich
introduces the chicken/egg other-way-around question, i.e. just what influenced what? :

From: Berj N. Ensanian KI3U
To: Journal of Voynich Studies
Sent: 12-26-2010 10:49:16 AM EST

J.VS: American Civil War coded message deciphered

Dear Colleagues

In this 25 DEC 2010 BBC News online article:

" Coded American Civil War message in bottle deciphered "

a picture of the message is provided, and we learn :

" A message in a bottle delivered to a Confederate general during the American Civil War has been deciphered,
147 years after it was written. In the encrypted message, a commander tells Gen John Pemberton that no
reinforcements are available to help him defend Vicksburg, Mississippi. "

" The text is dated 4 July 1863 - the day Vicksburg fell to Union forces. The small bottle was given to the
Museum of the Confederacy in Richmond, Virginia, by a former Confederate soldier in 1896. Earlier this year
the museum's collections manager, Catherine Wright, decided to investigate the wrapped note it contained. "

" When Ms Wright found that the message was coded, she asked retired CIA codebreaker David Gaddy to crack it
- which he did in several weeks. A Navy cryptologist later confirmed the interpretation. "

It would appear that if it took those fellows that long to crack the message, then it was strong enough to serve its
military tactical purpose.

Berj / KI3U


J.VS Archive continued in Vol. V, 2011