Category Archives: The Voynich Alphabet

Investigations of the shapes that are used within the Voynich to render textlike material.

The Pitfalls of Unhinging Pairs

An article posted today by Marco Ponzi alerted me to some VMS text analyses I have never seen before. A Google search revealed that several researchers have studied vowel placement in natural languages and some of those concepts were used to analyze VMS text (e.g., W.F. Bennett and linguist Jacques Guy).

So I looked up Guy’s papers, read through the first one, and am stating right up front that no matter how often one runs the numbers, vowel-identification analyses on individual glyphs in the VMS is not going to work as it does for natural languages. In this blog, I’ll explain why.


For those interested in computational attacks and historical precedents, here is a summary of my Google search for the VMS-related work of Jacques Guy:

  • Cryptologia 1991 (Issue 3), “Statistical Properties of Two Folios of the Voynich Manuscript” by Jacques B. M. Guy, in which he analyzes folios 79v and 80r as to letter frequency in terms of both word and line placement, and co-occurrence, with a “tentative phonetic categorization of letters into vowels and consonants using Sukhotin’s algorithm”. This was republished online in June 2010.
  • Cryptologia 1992 (Issue 2), “The Application of Sukhotin’s Algorithm to Certain Non-English Languages” by George T. Sassoon, in which he applies Sukhotin’s vowel-finding algorithm to a number of languages, including Goergian, Croatian, and Hebrew, et al. This was republished online in June 2010.
  • Cryptologia 1992 (Issue 3), “A Comparison of Vowel Identification Methods” by Caxton C. Foster, in which he compares four methods of vowel identification. This was republished online in June 2010.
  • Cryptologia 1997 (Issue 1), “The Distribution of Signs c and o in the Voynich manuscript: Evidence for a Real Language?” by Jacques B. M. Guy, in which he references Currier A and B and builds on the Sukhotin identification of c and o as vowels and speculates as to whether they might represent o and e, “which they resemble in shape”, “a phenomenon similar to that shown by Standard English and Scots English…”. This was republished online in June 2010.

There isn’t room to analyze all these papers in a single blog, so I will confine this article to vowel identification. Sukhotin’s algorithm is given by Guy in the first paper as follows:

“Given a text in a supposed unknown language written in some alphabetical system, Sukhotin’s algorithm identifies which symbols of the alphabet are likely to denote consonants and which vowels.”

Guy then illustrates glyph-assignments in W. F. Bennett’s transcription system from the 1970s from which his own glyph-assignments are derived:

As you can see, it’s very similar to the current EVA system.

Guy then illustrates his transcription system, which is very similar to Bennett’s, except the VMS double-cee shape is acknowledged, and some ligature-like VMS glyphs are transcribed as single characters (a decision that can be debated but may not have a significant impact on vowel-analysis statistics):

Is the Above Research Based on a Flawed Premise?

Guy’s research into VMS vowels, and that of many others, neglects important co-relationships in VMS glyphs. It’s not enough to come up with a transcription alphabet and then run frequency analysis software on individual glyphs (whether it is specific to vowels or not) if there is evidence of biglyphs or positional dependence.

Many researchers accept the spaces in the VMS as word boundaries and, for the most part, assume 1-to-1 correspondences between alphabetical letters and glyphs perceived to be vowels (either by humans or by the software). Guy does not acknowledge biglyphs in any significant way other than VMS “cc”, which has a historical precedent in early Medieval Latin texts of representing the vowel “a”.

Side note: In early medieval text, “cc” usually stood for “a”, but if “cc” was superscripted, the vowel “u”  was usually intended. The “cc”=”u” is not mentioned in Guy’s paper, he only references “a” and may not have known medieval Latin well enough to know that “cc” could also stand for “u” or the consonant “t”. He does note, on page 210, that what we call EVA-e “is always followed by {t}” but he’s referring to EVA=ch, not the “c” shape in general, which is followed by a number of glyphs.

Guy must have been working from a flawed transcript because he acknowledges “cc” and “ccc” but completely neglects “cccc”, a pattern that unambiguously occurs a number of times in the VMS, but which is almost entirely ignored in most transcripts (including the popular Takahashi transcript). But even this may not be a significant hindrance to analyzing text patterns.

A perplexing statement in Guy’s paper is that he acknowledges the ligature-like nature of some of the VMS glyphs but then partially negates this by stating: “It is certain, then, that {ct} and {et} represent single letters.” By “ct” and “et”, he is referring to EVA-ch and EVA-sh and I think there are many who would debate the certainty of this assertion.

Pairing-Patterns in VMS Text

I have described glyph co-relationships in previous blogs, including biglyphs, Janus pairs, positional constraints, and the “rules” for reproducing samples of VMS text from a conceptual basis and with examples, but this time I have decided to keep it simple and use just one glyph to get the message across in a more explicit way.

As an example of co-relationships that could significantly alter the results of computational attacks, I will refer to the backleaning glyph expressed in most transcripts as the Latin letter “i” (this glyph also appears on the third line of folio 116v in the middle of a word that resembles Latin “vix”, but this appears to be an exception).

Guy analyzed two different transcripts and, in his results charts, identifies the “i” glyph as vowel #6.

Here is a detail-clip of Guy’s results from one of the transcripts he used for analysis. The chart was published in Cryptologia, 1991, Issue 3, p. 211. Glyphs that have been statistically analyzed as candidates for vowels are listed in the left-hand column. You can see quite clearly that EVA-i appears only in the medial position:

The Crux of the Matter

The problem with analyzing (or perceiving) the “i”-glyph as a vowel (or any individual letter) is that it is only positioned in one way in the VMS.

This is in stark contrast to natural languages, where vowels are found in many positions, both 1) in relation to other letters and 2) within a word.

I call EVA-i the “pivot” glyph due to its position in tokens and the way specific glyphs precede or follow it and I have not been able to find any natural language in which any vowel is always preceded by the same letter, and is always in the medial position, as in the VMS. This is why I constantly refer to the “rule-based” and “positionally constrained” nature of VMS text.

Looking More Closely at the Rules for EVA-i

With the exception of folio 116v (which may be marginalia in another hand), the backleaning-i cannot stand alone and must be preceded by “a”. It can only be followed by certain specific glyphs or glyph-groups.

VMS “i” differs from “o” and “a” (identified as vowels #1 and 3 in Guy’s chart) in that “o” and “a” can be paired with other glyphs and can move around somewhat (and can appear at the beginnings of tokens), whereas the “i” glyph cannot.

Before one argues that “i” can be preceded by another “i”, consider this…

The minims that appear after “ai” in “dain” are commonly transcribed as “n”, “m”, “v”, and there are many places in the VMS where the scribes have written them with a curved connector similar to the shape “u” (in contrast to most glyphs in the VMS, which are not connected), so the general feeling is that the minims that follow “ai” are not necessarily additional “i” shapes. Guy also acknowledges this ambiguity on page 210, and Bennett  transcribes the minims as “m”, “n” or “u”.

After going through every glyph in the VMS numerous times when I was creating my transcripts, I record them as follows:

The glyph-pair “ai” occurs almost 6,000 times in the VMS. If you run the numbers on popular transcripts, you may get different results, because the distinction between minims is not recognized in many of them (as discussed in the above chart).

But even if the minims are interpreted in a completely different way from the way I have drawn them above, even if one supposes all the minims are the same glyphs, the essential problem remains… the minims, as a group, must be preceded with “a” and do not occur at the beginnings of tokens, only together, and only at the end. The argument for “i” being a vowel becomes even weaker!

Guy acknowledges the possibility that the “i” strokes might be minims (or something else), but nowhere does he address the important fact that however one transcribes them, individually or as a group, “a” must precede them and they are never at the beginnings of words. These characteristics, taken together, are why we must question their interpretation as vowels.


Early researchers like Bennet and Guy were working with low-resolution B&W photostats, so some of their misconceptions can be forgiven, but ignoring glyph-placement is hard to excuse. Even if you can’t see fine details of individual letters, there’s no mistaking the following properties of EVA-i:

  • EVA-i is virtually always preceded by “a” (there are 7 rare instances (only 1/10th of 1%) of an “r”being inserted between two minims and one is especially strange as it is in different handwriting, is out of line, and is proportioned differently),
  • EVA-i never appears at the beginnings of words,
  • EVA-i rarely appears at the ends of words unless all the minims are assumed to be the same, and then they are always at the end, and
  • only certain specific glyphs follow “i” (I’ll include further details in a future blog).

The only way to resolve “i” into a natural-language vowel is to consider additional ways in which the text might be manipulated (as examples one would have to manipulate spaces, letter-order, or assign the same letter to multiple glyphs, or perhaps assume there are “hidden” minims, which is starting to stretch things a bit far, etc.). As it stands, if the spaces and letter-arrangement are taken as literal, and the glyph-assignments are considered consistent throughout the document, there’s not much evidence to support EVA-i as a vowel using the form of analysis proposed by Guy.

One might try to relate ai to “qu” in the sense that they are often found together and “q” rarely exists without “u”, but “qu” occurs in many different locations in a word, as can most common pairs of letters in most languages. The analogy doesn’t hold.

The Consequences

The behavior of EVA-i also affects the interpretation of EVA-a. If “ai” turns out to be a biglyph, then other instances of “a” have to be evaluated separately from “a” + “i” and the statistics will change, as will the number of glyphs that form the Voynichese “alphabet”. If there are other biglyphs (which I believe there are, as described in my blog about Janus pairs), then all the single-character computational attacks and early “vowel-assignment” research needs to be re-visited.

If, on the other hand, “ai” is not a biglyph, EVA-i is still problematic because one has to ask, “What is the purpose of the preceding “a”? or of “i” itself?” It’s possible that neither EVA-i, nor some instances of EVA-a, are vowels. They might not even be letters, but even if they are, they can not be statistically evaluated without taking into consideration the relationship of “i” to its companion.

J.K. Petersen

Copyright © 2018 J.K. Petersen, All Rights Reserved

Latin’s “Om-age” to Indic Numerals

Most people don’t think of Indic and Latin scripts as similar, but the links between east and west are old and deep and medieval Latin script is not the same as modern Latin.

When I first discovered VMS glyphs, I scoured foreign alphabets for the origins of some of the less familiar characters. I already knew the Latin alphabet, some of the runic scripts, the Cyrillic and Hebrew alphabets, the rudiments of Korean, a little bit of Russian and Japanese (and a tiny bit of Chinese), some Coptic Greek, a few Greek numeral systems, and a smattering of Malaysian alphabets, but no matter how hard I searched, none of them, except Latin (combined with a small percentage of Greek), seemed to match a high proportion of the VMS glyphs.

I also searched plant-related words in Baltic and Turkic languages. Unfortunately, I haven’t had time to study Finnish, Czech, or Silesian, but they’re on my list.

Just to be sure I hadn’t missed anything, I explored several other alphabets from languages I thought had potential, including Georgian, Armenian, Amharic/Ge’ez, Syrian, and Sanskrit/Gujarati/Nagari (the word Devanagari did not exist in the middle ages) and… once again was led back to Latin, but with a better understanding of how Latin, Greek, and Indic script were more similar in the Middle Ages than they are now.

Western Presence in Eastern Lands

In ancient times, the Greeks and Romans occupied Pakistan and made forays into northern India. Alexander the Great, the Kushana peoples, and the Persians all left their mark, and absorbed certain aspects of Indic culture. There were numerous Indic coins that included Greek letters and numbers long after Greek occupation had subsided.

I couldn’t help noticing that “Arabic” numerals, as they were used by Latin scribes in the 14th and 15th centuries, resemble Indic numerals more than Arabic, and I subsequently saw the credit line in Latin, in the Codex Vigilanus (Spain, 976), attributing the number system to the Indians.

The earliest-known Indian numerals in a European manuscript are in the Codex Vigilanus (976 CE). It’s possible the manuscript reached the Spaniards through Arabic traders, thus leading to the “Arabic” moniker.

Leonardo of Pisa, now known as Fibonacci, appears to have independently discovered the Indic number system that was documented in Spain two centuries earlier. While traveling in Bugia, North Africa, with his father, he observed the notation system and calculations used by Muslim traders. When he returned to Pisa, he wrote Liber abbaci “Book of Calculation”, which included the Indic numerals. There are no copies of the original, completed in 1202, but a number of copies of Fibonacci’s enlarged 1228 edition survive.

The following is from a copy of Fibonacci’s book, believed to be from the late 13th century (BAV Pal. Lat. 1343). Like the Spanish manuscript, it introduced the numeral system that became popular until the 15th century, when slightly rotated glyphs for 4 and 7 and a more curled 5 evolved into our modern system:

Despite widespread acclaim for Fibonacci’s 13th-century manuscript on computation, change occurred slowly, and Roman numerals did not significantly give way until the 15th century when more flexible calculations were needed for scientific studies.


Latin Conventions in Medieval Scripts

Researchers often miss similarities between VMS glyphs and Latin because medieval scribes used many ligatures and abbreviations that are not taught in modern Latin. These were as integral as the letters themselves, and it’s hard to find late-medieval manuscripts without them.

Before describing similarities between Latin and Indic scripts, it’s important to understand how Latin is more than just an alphabet. You’ll note in the examples that follow that several of these scribal conventions are apparent in Voynichese.

Example #1

The first sample (BNF Lat 731) is lightly abbreviated. It uses some of the more common Latin conventions, including quibus, per, et, tails on the ends of words that loop back over the previous letters to indicate missing letters (it’s like an attached apostrophe), and caps over other letters to serve much the same purpose when the missing letters are closer to the middle of the word than the end.

Notice that loop-back tails and caps are common in the VMS, and that the abbreviation symbol that resembles a “2” or back-leaning “r” is, as well.

Example #2

The second example (BSB CLM 29505) also uses very common conventions, but not identical to the previous example. Scribes were free to pick and choose what was convenient because they were interpreted by context.

In this example, we see the common symbol for “Item” (at the beginnings of lines)—it resembles EVA-k; the macron or “cap” that indicates missing letters; the swooped-back tail at the ends of words (also missing letters); g° to stand for degree (grado/grade); a squiggly line over the “e”, which usually indicates a missing “r” or “er” “ir” or “re” (again, depending on context). Note that this is similar to the squiggle on the red weirdo on VMS 1r.

The loop on “item” is also used at the ends of words to represent “is” with the Latin suffixes -ris/-cis/-tis being drawn like EVA-m.

Notice also the tail on the “r” on the last line. This tail wasn’t always added to “r”, sometimes it was added to “i”, so one has to read for context to know which letter was intended. Take note that the shape of the tail sometimes indicates specifically which letters are missing (I’ll come back to that later), but not all scribes distinguished the missing letters by shape.

Thus, there are four scribal conventions in this small sample that are found as VMS glyphs:

Example #3

The third example (Ms San 827) makes slightly more frequent use of abbreviations, but they are still very common ones and easily readable.

In sample #3, note the lines and caps over the letters to indicate missing letters, the curled tail on the “p” to stand for “pro”, the symbol that resembles a “2” which sometimes means “et” (and) but often means -ur or tur.

On the fourth and fifth lines, you will see the “9” symbol at the beginning of one word and the end of another. At the beginning, in this example, it stands for “con-“. At the end it is usually “-us” or “-um”. This is one of the most common glyph-shapes in the VMS and, as in Latin, it is usually at the end, but sometimes at the beginning:

Example #4

The above examples are all from the 15th century, but conventions were similar in the 11th to 14th centuries, leading up to the creation of the VMS. The following earlier text (OBV SG 21), uses all of the same concepts and most of the same conventions:

Thus, with four brief samples, and the numerals that evolved from Greek that were mentioned in a previous blog, we can account for the majority of glyphs in the VMS.

The problem is not in relating the VMS glyph-shapes to Latin letters, ligatures, and abbreviations—the similarities are numerous and obvious—the difficulty is in determining their meaning because VMS tokens do not, in general, behave like Latin or the majority of natural languages in terms of the variability of the words or the characters within the words. Here are some important differences:

  • In Latin scripts used for a variety of languages, abbreviation symbols can be associated with many different letters. In the VMS we see caps only on EVA-sh and occasionally EVA-q.
  • In Latin, the swept-back tail is found on almost any character where letters have been omitted near the end of a word. In the VMS, it is specific to EVA-e, EVA-r, and the last glyph in “daiin”.
  • The “9” symbol is shaped and positioned the same in both Latin and Voynichese, but in Voynichese it’s much too frequent to mean the same thing as it means in Latin (or other common languages).

So the shapes are similar to Latin, but the extreme repetition and positional rigidity are not.

After the 15th century, abbreviations and ligatures fell out of use, as Latin scholarship was replaced by local languages, and the newly invented printing press and typewriter introduced mechanical limitations that made it difficult to mimic these scribal traditions.

Ties with the Eastern World

So what does all this have to do with the Indian scripts mentioned at the beginning?

Dozens of languages have been mentioned in connection with the VMS, but claiming it’s a specific language is easy. I saw one person claim five different languages in the same week, and another claimed three more in the course of three months. Proving that it’s a specific language is the real challenge, and so far no one has provided a convincing translation of even one paragraph.

I think I know why so many different languages have been proposed for the VMS. It’s partly because expanding or anagraming text expressly turns it into readable text or, if Voynichese is based on natural language, it may be partly because words related to disciplines like science are often loanwords and thus similar in many languages. But this bewildering array of suggested languages might not be entirely imaginary… certain languages did, in fact, have more in common with one another in the Middle Ages than they do now.

As an example, Indo-Iranian writing styles are more similar to medieval Latin than east-Asian character-based scripts like Chinese—both come from proto-Indo-European roots.

The Indo-Greeks and others who subsequently ruled Pakistan kept some of their native customs and adapted others from local culture. They blended pagan gods with Buddhist beliefs and minted bilingual Indo-Greek coins, as in the following example from c. 100 BCE:

[Image courtesy of the Classical Numismatic Group, Inc.]

The Kushana, nomadic peoples from central Asia, at one time ruled a large region that included Afghanistan, parts of Pakistan, and northern India, and almost shared a border with the Romans during Trajan’s and Hadrian’s rules (a coin mould featuring Emperor Hadrian was found in excavations of c. 2 CE artifacts in Rairh, near New Delhi). The Kushana were Indo-Europeans who actively traded with both Rome and China.

This gold coin, probably of Kushan origin, is a testament to multicultural interaction. It was minted in India, inscribed with Greek letters with the ruler on one side and “Boddo” (Buddha) on the other, and was unearthed in Afghanistan. Sometimes Zeus was substituted for Buddha on this style of coin.

[Image courtesy of the British Museum.]

Commonalities with Indo-Iranian Scripts

Please note that I have used Gujarati as an example of glyph similarities, even though it is more recent than Nagari, because it does not have the line across the top (thus making it easier for westerners to read). It is very similar to other Indic scripts if you ignore the top-line and look specifically at the shapes underneath. The following observations apply to a group of related Indic scripts descended from Sanskrit, not specifically to Gujarati.

I’ll start with some of the simpler and more familiar shapes, followed by glyphs with ascenders (gallows characters), because the majority of VMS glyphs are Latin. Only a few that are rarely used (or which show up only once) are distinctly eastern and will be described later.


Glyphs with Tails

Voynichese has a number of glyphs with tails, a ubiquitous convention in medieval Latin. Adding a tail to a glyph wasn’t just an embellishment, it was a way to indicate missing letters. In the VMS, the r, c, and minim shapes at the end of the word “daiin” all have distinctive tails. Certain Indic glyphs also have tails, and the shape or length of the tail can change the sound or meaning of a letter.

Here are some interesting patterns in Latin and certain Indic scripts, that may have some relevance to the VMS:

  • EVA-r. In Latin, when a tail is added to “r”, it can mean “rus, but it often means “re”, “er”, “ra”, “ar”, “ir”, or “ri”. In other words, a vowel is inherently indicated by a tail added to a consonant, as in some of the abugida languages. Similarly, in the later 13th- and 14th-century Nagari scripts, and in Gujarati, you will see an “r” shape with a curved tail to represent “r” or “ar” or “ra”. There are several places in the VMS where two forms of tails are apparent in the same block of text. In Voynichese, Latin, and Gujarati, the curved tail is more frequent than the extended-loop tail. If Voynichese is anything like Latin, Gujarati, or some of the Malaysian scripts, and not just a smokescreen to make the text look like Latin, then extending the tail and changing its shape changes the meaning of the glyph:
  • EVA-s. In many older Latin scripts, the “t” was written like a “c”, rather than with a straight stem. It can be a struggle to tell them apart. Adding a tail to this c-like tee stood for “te” or “ta” or most combinations of “t” plus a vowel (it can also mean “ter” or “tus”). In Gujarati, the symbol for “ta” is a c with a tail (note that both “r” and “c” shapes with tails are found in the VMS) and some are ambiguous, with a slight hook on the foot, perhaps denoting a third character. In Greek, a c-shape was used as an abbreviation for “kai” (and). Once again, if you look at it from a Latin point of view, the c-shape can also be “e” (many early medieval e-shapes didn’t have a crossbar or hook), and adding the tail turns it into “eius” or “et” for “and” (in fact, if you extend the tail a little more, it becomes an ampersand). Thus, we have a glyph with many meanings. C-tail can be the abbreviation for te, ta, or ter, or for et, eius, or er. In the VMS, as in Latin, this tailed shape, which sometimes resembles c-tail, sometimes e-tail, and sometimes t-tail, is found both individually and within other words.
  • EVA-d. If you look at variations of the thorn character, which is usually associated with northern European scripts, you’ll see some of them are written like a curvy “d” or a Greek sigma with a small bar through the ascender. It may be coincidental, but the Gujarati shape for “tha” is a curvy “d” shape. There’s no line through the stem, but many Latin scribes wrote it that way, and there is a strong association between “d” and “th” sounds in various Indo-European languages. If you round the top loop a little farther, as some scribes did with Latin “d”, thorn, and Greek sigma, it resembles a figure-8. This is why many researchers read the figure-8 on folio 116v as an “s” or “d”, but perhaps “th” should also be considered.

There are analogs to VMS shapes in both medieval Latin and some of the Indic scripts. The “a” and “o” shapes need no explanation—they are distinctly Latin, and “o” is common to many languages.

The simple “c” shape doesn’t tell us much either, because it is found in most alphabets, but two c-shapes tightly joined were used in early-medieval Latin to express “a”, “t”, and sometimes “u”. The double-c is also found in the VMS (right)—a distinction that might be meaningful but is not recognized in most VMS transcripts. In fact, in the Takahashi transcript, which is probably the most widely used, the extra c-shapes are sometimes omitted.

But tails are meaningful in both Latin and Indic languages, and ligatures common to both. Sometimes the tail changes the letter, sometimes it extends a sound, and sometimes it specifies which vowel is used. Note that Nagari and Gujarati are syllabic languages which might not seem to have much in common with Latin, but medieval Latin script has its share of implied vowels.

A sidenote on abugida scripts… Gujarati is a syllabic language, but not entirely an abugida script (neither is Hebrew). Both Hebrew and Gujarati include a shape for alpha, so it is explicit rather than implied (it’s possible that in ancient languages alpha was more of a glottal stop than a vowel), but most of the time the most common vowel (alpha) is rolled in with the consonant, as it is in a number of Asian and African languages.

In Gujarati, several of the syllables are written as though they were ligatures, with a vertical stem on the right  (as in sa, pa, na, and numerous other glyphs). This is technically part of the syllable but can also be thought of as the implied vowel. This vertical line has an additional function—it can be added to the preceding vowel or syllable to lengthen it into a long vowel, as in the following example:

Note how the vertical bar changes a short-a to long-a, a symbolic concept that was mentioned in the previous relative notation blog. A similar convention exists in Modi, another Indic script that is first recorded in the late 14th century.

Some of the commonalities between Latin and Indic scripts disappeared when Latin abbreviations were dropped and Latin was reduced to a simple alphabet.


I have much more information on this subject and was going to try to cover the Voynchese ascenders and some of the rare characters in the same blog  because they also have their roots in scribal conventions, but this is becoming too long, so I will continue with the less common characters in a future installment.

… to be continued…

J.K. Petersen


© Copyright 2017 J.K. Petersen, All Rights Reserved

A Helping Hand

What Can Hands Tell Us?

When I was creating a VMS transcript, I noticed immediately the change in handwriting partway through the big-plants section on folio 26r. It was not just scribal haste or fatigue—the spacing, rhythm, and slant were different, as were some of the letter forms. It was clearly the same style of writing (perhaps a blood relation of the original scribe?), but not the same hand.

Interpretation of VMS glyphs is something I’ve wanted to write up for a long time but I wasn’t sure how to express it in a way that was sufficiently clear, until I realized the hand on folio 95v1 might help me illustrate the concepts.

In the following illustration, there are several instances of EVA-d with a straight stem (marked in red), and certain glyphs with greater separations between their component shapes (marked in blue):

I have often wondered whether a rounded “d” and a straight “d” are different glyphs, and created two different characters in my transcript to record them. But I still treat them the same most of the time, as they seem to fall into similar patterns. But perhaps they are different. For example, when they are at the ends of words (which happens frequently when the d is paired with EVA-y), maybe the two shapes are meant to be read as two different endings. If this were Latin, for example, one might mean -us and the other -um, or one might mean -us or -um and the other might mean -bus. Or maybe there’s a completely different interpretation (that I’ll discuss later).

Another thing I noticed on this folio is the greater tendency of the scribe to separate the component shapes of a glyph. There’s a good example on the far right (marked with a red arrow) in which the first curve is clearly separated from the “is” shape (“is” is a Latin symbol that looks like a short cursive ell). The “is” shape occurs in EVA-m and gallows characters and sometimes the letters with “is” are written almost like short gallows, suggesting they might be related.

What Do Tails Tell?

There’s another distinctive aspect of Voynichese that inspired me to create my own transcript and my own fonts. Notice how strongly the character normally referenced as “n” (in daiin) resembles a v or a w? This is how I transcribed them. But then how can you tell the difference between v or w if there are one, two, or more minims preceding them? This is something I pondered for a long time and I think the answer (at least for this scribe) might be the length of the tail. Notice how the tail loops back farther on the one that resembles “w”. I don’t know whether v and w are meant to represent two different characters, but I think the distinction between “n” and “v” in the transcript is important, as I’ll explain farther along.

Enumerating the Gallows

Some years ago, when I was looking up the history of pilcrows and gathering samples (which took a couple of years), I also collected examples of Greek and Latin abbreviations and number systems because many of them resemble gallows characters.

Early on I was insisting that almost all the VMS characters are based on Latin (with a few on Greek) and there was a lot of resistance to the idea (I got some “interesting” email). Quite a number of people disagreed with me, some rather disparagingly, and said I should be looking at Armenian or Georgian, or other script systems dissimilar to Latin because, as they said, “It doesn’t look anything like Latin.”

I have looked at those other alphabets (and many Asian scripts, as well) and still maintain that the majority are based on Latin character-shapes and abbreviation conventions, as I’ve noted in my blogs. But maybe things are changing. I’ve noticed a recent upswing in VMS “solutions” claiming that the text is Latin that needs to be expanded. Well, maybe, but I want to emphasize the fact that Latin characters and scribal abbreviations were used in many languages, not just Latin, so Latin glyph-shapes don’t automatically mean Latin language.


But to get back to similarities with Latin abbreviations, a horizontal line or slightly slanted line was commonly used in early Latin documents to signal missing letters (similar to an apostrophe). Here are some examples of abbreviations and ligatures (which are not the same thing and should not be confused with one another):

And now we get to the good part…

If you look at the first illustration again, where separations between individual parts of a glyph are more distinct, you might notice the Vword on the bottom-right, usually transcribed as “dal” looks suspiciously like the Roman numeral dcix (I was tired when I wrote this, this is 609, not 59). In my transcript, I have transcribed EVA-l as “x” for the simple reason that it looks more like a medieval “x” than “l” to me, but also because I noticed the similarity between Voynichese and Roman numerals early on and wondered whether there might be a connection.

Greek and Latin Numerals and Their Relationship to Voynichese

Old forms of Greek, Hebrew, and Roman script did not have a separate set of glyphs to express numbers. Instead, they were written with letters. Over the centuries various conventions were used to mark them so they were not mistaken for letters.

In Greek, a line was drawn above or through the character to signify a number. In Latin documents that used Greek conventions, some numbers were expressed using Greek forms, some were in Roman numerals (sometimes with a line over them), and some were Arabic.

Here are examples of numbers from Greek and Latin manuscripts that may have inspired the benched gallows characters in Voynichese. Note also that if you’re not a paleographer and you came across the Greek examples (top row), without the Latin examples I’ve added below them for comparison, you might be mystified as to their meaning:

Look at this excerpt from 95v1 one more time, paying particular attention to the characters in the bottom right. Note how the separated “a” glyph makes the token look like dcix (609) in Roman numerals.

In fact, the text directly preceding “dax” looks very much like Mccdciiiv, which isn’t quite conventional, as two would usually be indicated with “ii” rather than “iiiv” and you wouldn’t normally place a dee between the three cees, but what if the tail is the common Latin abbreviation for a line over the letters, which was sometimes written as an attached tail to facilitate quick writing? Then you get Mccdciiii—still not quite conventional, there’s still the problem with the ccdc, but notice that the cc is benched.

Hmmm, could the bench on the cc (EVA-ch)… possibly mean it belongs on either side of the preceding “M”? Maybe what we are looking at is cMɔ dciiii (with a tail over the iiii to indicate a number), as it is in the illustration above. This can also be written with a pipe symbol as follows: c|ɔ dciiii as it was often written in the 15th century and onward.

The d-“aiin” token comes in many flavors. It’s not always preceded by “d”, it can be preceded by almost anything, and the number of minims after the “a” shape ranges from 0 to 4. If the stem of the “a” is also a minim (if a is ligature c + i), then it ranges from 0 to 5 or 0 to 4 plus “v” (Roman numeral 5) depending on whether one interprets that last glyph as a “v” or as an “i”-with-a-tail to indicate a number.

Inspiration for Shape and Structure

Is Voynichese numbers? If it’s numbers, do they represent letters or sounds? Or is Voynichese a coding system that includes a subsystem for numbers?

If you take your mind out of linguistic mode for a few moments, and pretend the text is Roman numerals (even if it isn’t), do you notice that you see it differently? Have you made assumptions you didn’t realize you were making?


As I’ve posted in many blogs, the glyphs are based on Latin letters and abbreviations, but they look to me like they’re based on specific Latin characters that have a high correlation to Roman numerals.

Roman numerals consist of M d c l x v i (sometimes scribes lined up several “i” characters instead of combining them with v), sometimes c-shapes were placed on either side of the M, sometimes a line was drawn over or through the letters. All of these are hauntingly similar to aspects of Voynichese.

Notice how the characters that are benched resemble tau and rho, the two characters placed above “m” (which was sometimes written as a bench in both Greek and Latin). In Greek, a rho looks like a “p”.

Even the EVA-r glyph might not be an “r”—it might be an “i” with a tail (as it was written in Latin).

Except for EVA-o, -y, which are suspiciously frequent compared to natural language frequencies, and EVA-q, which is very positionally consistent, almost all the common VMS glyphs bear a strong resemblance to M d c x v i and benched forms of  tau, rho and M + c. Note also that EVA-o and y are variations on circles. Maybe EVA-o, y, and q are some kind of markers.

Or maybe “o” stands for zero (as in 1408) or is another form of “c” (depending on position).

The Voynich characters are positionally constrained. So are Roman numerals. If you put a “d” in front of a “c” it means something different from “c” in front of “d” (600 versus 400). Maybe Voynichese does this too.


The VMS might not be Roman numerals, it might not be numbers, but there is a strong similarity between Voynichese glyphs and Roman numerals.

There is also a strong similarity in how Voynichese prioritizes glyph order. Whatever system is behind the VMS, I think Roman numerals, at least on the conceptual level, had something to do with the way Voynichese was designed.

Perhaps other people have mentioned Roman numerals in connection with the Voynich Manuscript, I don’t know (it seems like a reasonable supposition and I’m still comparatively new to the Voynich scene), but I haven’t seen anyone demonstrate a connection between benched numbers (Greco-Roman glyph conventions) and the bench characters in the VMS. Nor have I seen anyone provide a cohesive explanation of how VMS glyphs may have been historically and pictorially inspired by a system like Roman numerals, so hopefully this will add something new to the VMS corpus.


Before I close, I have a little bonus… It’s a secret where I found this (at least for now), but here’s a little medieval “pen test” that you might enjoy.

Note that only three characters are needed to represent the whole alphabet, except that one might need a few nulls to separate the individual “letters” and to obscure the fact that they are Roman numerals so it won’t be too easy to break. Imagine what it would look like if you did that?


J.K. Petersen

© 2017 Copyright J. K. Petersen, All Rights Reserved



Janus Pairs

The Theme of Duality

You may have noticed pairing in the VMS… double crayfish, two sheets each of “Aries” and “Taurus”, 2 x 2 sets of 17 on the page that looks like a code wheel, but have you looked for pairing in the text?

In a previous post I introduced a group of glyphs I call The Gang of Four (if you haven’t read it, I strongly suggest it or this followup blog won’t make much sense). The Gang of Four is a subgroup of glyph-pairs that occur with great frequency within Vwords, and can also stand alone. Together with other glyphs with similar properties, I refer to them as Janus Pairs (or JPairs for short).

Janus was the god of duality. He presided over beginnings and endings, doorways and passages. I like the analogy of Janus Pairs opening a window into the structure of the text.

Unlike English and many other languages, VMS glyphs cannot be shuffled in a multitude of ways to create a large pool of Vwords. Certain glyphs are found only in certain positions. This is true even if you evaluate them in pairs, which means the VMS is more positionally rigid than syllabic Asian languages, as well.

But pairs there are, and they form a disproportionately high percentage of letter combinations in Voynichese, with some interesting differences in where they are used.

The Prevalence of Janus Pairs

I cannot fully describe the dynamics of Janus Pairs in one blog or two, any more than one could describe the dynamics of English in one blog or two, but I can introduce them so you can visualize the patterns and make sense of follow-up articles.

Before posting examples, please be aware that I’ve spent years trying to discern which are pairs and which are monoglyphs (or ligatures). This is not easy (if it were, it would have been done a long time ago), but some can be confirmed by following them through the entire manuscript and noting where they fall in relation to other glyphs. It took me a while to figure out how to present them so the patterns could be readily seen.

The Gang of Four is an example of prevalent pairs that can be either free-standing or joined to other words, but it’s important to look at all the JPairs. Unfortunately, it’s not practical to post all of them, so I have selected examples from two sections.


The first group is from the “zodiac” section. There isn’t room for all the zodiac symbols, so I selected four as examples and chose only the text from the labels (not from the text inside the double rings). If this subject interests you, you can look at transcripts yourself to work out the others.

The second group is from the big-plants section. I’ve chosen two plants near the beginning, and one farther along.

Obviously, to understand the text, you have to analyze and compare all sections and all the Vords on each page, and I have spent years doing this and still have some unanswered questions, but the following charts should be enough to get the concepts across. Note that I have chopped three of the less common Pisces labels from the bottom of the chart, mostly to save space, but also to put the emphasis on the ones that are most prevalent and most illustrative of patterns. This doesn’t mean the three deleted Vords are unimportant.

I have color-coded pairs to make them stand out because I’ve seen so many decryption attempts that don’t take them into account. These charts are not designed to reveal the meaning behind the text, that is best done by organizing them in several different ways and placing them side-by-side on a very long table. Their purpose is to illustrate

  • fundamental positional patterns,
  • pair composition,
  • pair frequency,
  • order of glyphs within pairs,
  • and differences between the text in two different sections of the manuscript.

So here is the first set of Janus Pairs, from the zodiac section:

I’ve collected samples of text in a number of languages to compare to these patterns (which is another long subject, possibly too long for a blog).

Here is the second chart, with examples from the big-plants section (note that a few Vords are chopped from the bottom of the first column due to space constraints).

You can immediately see that they differ in form and content from the zodiac-symbol vords but that there are structural similarities in where glyphs appear in specific vords (note that I am not certain aj is a Janus Pair, it can sometimes be oj and may be two separate glyphs).


Even though these are only small excerpts, there is much information to be gleaned from them.

Note the overall differences between zodiac Vords and plant Vords. There is a high prevalence of ot and ok combinations in the zodiacs. In the plant section, one sees many Vords starting with EVA-ch, -sh, or d and few of those that are common in zodiacs. If EVA-ch is a ligature then it may also qualify as a pair.

These patterns are prevalent enough that it’s possible to make a few predictions about where vords are likely to appear in the manuscript. You can’t do it from these charts alone (although some of the patterns are more obvious than others), but it’s possible when all the tokens in the manuscript are evaluated together.

Note also that there are priorities in terms of glyph placement. An “o” glyph paired with EVA-l, -t or -r, for example, behaves differently from one that isn’t combined into a pair group. This might be one of the reasons the o-glyph is so frequent, and also suggests that some glyphs may be intended as monoglyphs or ligatures—their status may be determined by their position and relationships to other glyphs, which might explain the strict positional rules.

VMS text is highly structured, not at all random, and there is substructure within individual sections. As to how it relates to natural languages, I’ll discuss that in a future blog, after you’ve had time to digest this one.


J.K. Petersen

© Copyright 2017 J.K. Petersen, All Rights Reserved


The Gang of Four

I’ve wanted to blog about VMS biglyphs for years, and have alluded to it in several blogs, but simply couldn’t figure out a lucid way to illustrate the patterns. Recently, I came up with an idea that might make it easier to explain.

Some Brief Background

I’ve already written about how the EVA-y glyph appears frequently at the ends of Vwords and sometimes at the beginning, a pattern very prevalent in medieval Latin. In Latin, this glyph is based on the number 9 (to distinguish it from the letter g) and usually represents -um or -us at the ends of words and con- or com- at the beginnings of words (see example right). Thus, a single glyph can be expanded in at least four ways, and its meaning known by context.

The apostrophe, shown here as a curved “cap”, similar to the cap in EVA-sh in the VMS, can also be written as a short line, a long line, or a squiggly line and can represent one, two, or many missing letters.

If Voynichese were meaningful (and somehow encrypted), and if some of the VMS glyphs are meant to be abbreviations, it would affect both frequency and entropy calculations and would not be readable using one-to-one substitution codes. Attempts to expand the abbreviations using software algorithms would be challenging, as well, if one considers that the medieval apostrophe could stand for almost anything, was not used consistently, and wasn’t always placed above the area where letters are missing.

Also, it’s important to keep in mind that Latin abbreviations were used in all major western languages, not just Latin, and their meaning adapted to common patterns for each specific language.

There have been a number of Voynich “solutions” lately that claim the text is abbreviated Latin (an idea that has been around for a long time). It’s important to keep in mind that Latin symbols do not automatically mean Latin language, just as Cyrillic characters don’t automatically mean Russian. Many languages are written in Cyrillic, including Mongolian, Bulgarian, and Ukrainian.

The Fearsome Foursome

In the process of trying to discern whether Voynichese is intended to be expanded and whether certain glyphs behave in specific ways that might reveal whether they are letters, abbreviations, or modifers/markers, I’ve been studying a group of glyphs that stand out as different.

Note that this article is not about abbreviations, it is about a set of glyphs I call The Gang of Four. The above note about abbreviations is a necessary preamble to explain why a fifth glyph-pattern that superficially looks like the other four doesn’t necessarily belong in the same group.

Also… all the following charts and numbers are based on my own VMS transcript, so there may be small statistical differences compared to other transcripts, but the overall concepts still apply.

First, before I go into detail, try this little experiment, it makes it easier to see the patterns.

• Take the two paragraphs on folio 1v.

• Do some search-and-replace and remove all the commas, spaces, line breaks (but not the paragraph breaks) so you have two long continuous lines of text. You should end up with something that looks like this (this is my “easy-read” VMS font but you can do this with a transcript character set or with the EVA Voynich font):

• Save a copy of the processed text so you can use it again, it sometimes takes a couple of go-rounds to get used to seeing the pairs.

• Now remove the following characters (I have specific reasons for choosing these characters): EVA-ch, EVA-sh, EVA-d, EVA-s, EVA-q, and whatever follows the “ai” in aiin or daiin (depending on the transcript, this may be one, two, or three characters), and EVA-q.

Now your text should look like this:

Take the beginnings of paragraphs with a grain of salt. There may or may not be pilcrows that behave differently from other glyphs depending on their position.

Starting after the first glyph in the first paragraph, walk through the text and add spaces so you are breaking it into pairs with the exception of “air” which is to be treated as a triglyph. Consider a benched-gallows to be a pair. You will notice the paragraph breaks fairly naturally into pairs except that there is an extra “o” once in a while.

Do the same thing for the second paragraph starting after EVA-Po (can you see why?). Again, treat “air” as a triglyph.

If you pay attention to the glyph pair patterns, you get something like this. Once again, it breaks down fairly naturally into pairs except that there are a few extra “o” glyphs (as in the first paragraph) and occasionally the gallows k or t stands alone.

These are the same pair patterns I pointed out in a previous blog but I realized later that I should have colorized them to make them easier to see:

I’m not sure of the significance of the extra “o” glyphs that sometimes occur between pairs, but I suspect that the o-glyph, when not paired might be a null or modifier (I am not certain of this, but there is a very high proportion of o-glyphs, and other glyphs like r or l or a do not show this propensity to appear in between common pairs).

Positional Flexibility and Doubled Letters

If you’ve studied the VMS glyphs individually, you’ve no doubt noticed that their positions are very constrained and that doubled letters are uncommon. And yet, even after removing seven glyphs, if one evaluates the processed text in terms of biglyphs (and perhaps a small number of triglyphs like “air”), then there are enough pairs to make a full alphabet. The peculiar lack of doubled letters in the VMS, and the positional rigidity changes when the text is evaluated this way.

I’m not suggesting this is a solution to the VMS or that the glyphs that were removed have no meaning. I’m using this as an exercise to focus eyes on certain important patterns that exist within the text that seem to be frequently overlooked and which change the dynamics of text breakdown and their statistical properties to a considerable extent.

So why did I choose seven specific glyphs to remove? Mostly to remove visual clutter to emphasize the glyph pairs, but also because I believe the ones that were removed may be ligatures (two shapes combined) and thus function as pairs on their own. That EVA-ch may be a ligature is suggested by its behavior and also by the gap that occurs between the left and right sides on folio 1r. Benched gallows characters are more obvious candidates for ligature-biglyphs and do appear to behave as such, so I left them in for this example.

Of the seven excised glyphs, EVA-y might be a special case. It doesn’t behave like the others. I strongly suspect it was added to make VMS superficially look like Latin and, of all the characters in the manuscript, if there ARE nulls, this one should definitely be considered.

Statistical Studies

If the VMS is constructed from biglyphs rather than monoglyphs, then many of the existing computational attacks would be irrelevant. I’ve been studying the biglyph-patterns almost since I first saw the VMS, but finding ways to describe their existence, their behavior, and especially their significance has been a challenge… which brings us back to the Gang of Four.

There are four biglyphs that form a statistical cluster and a couple that look superficially similar but behave a little differently. These biglyphs stand alone or act as part of other VMS tokens. Note, this is not a full chart of all two-glyph Vwords, there are several more, but these are ones that occur most frequently with spaces on either side and which can also be attached to other Vwords. Note also that if some of the deleted glyphs in the example above are confirmed to be ligatures, to represent two glyphs with one shape, then at some point, they must be evaluated in conjunction with these.

As can be seen from the chart, ox, or, ar, and ax cluster at the top in terms of how often they appear independently (with spaces on either side). They can be at the beginning, middle, or ends of Vords, indicating positional flexibility that is absent from monoglyphs when they are evaluated individually. I would have liked to include EVA-ot and -ai on the chart because they follow soon after those illustrated above, but for visual clarity, decided to exclude them for now.

The Voynichese snippet mixed in with the other text on folio 116v is from this group, as are many of the VMS labels.

The odd combination of EVA-dy, in the fifth position on the chart, is almost always at the ends of Vords, and with suspicious frequency, more than one would expect with natural languages. I am reserving judgment on this pair, but feel that it may be a null calculated to make the VMS text resemble Latin or a generic syllable intended to be interpreted in a variety of ways (and yet still calculated to look like Latin).

The second odd combination of EVA-am sometimes appears in several positions in Vords but is most often at the end and very frequently at the end of the line and thus behaves quite differently from the first four pairs and somewhat differently from EVA-dy. It is less often attached to other Vords than the previous five.

These patterns can be seen in many of the VMS labels.


It is tempting to think that The Gang of Four might be vowels, as vowels are the most commonly used letters in many languages. Vowels can sometimes stand alone (depending on the language) and could conceivably have been crafted for the VMS from four combinations of two glyph-shapes to make them easier to remember or recognize when writing or deciphering the text.

Testing this idea is harder that one might expect (which is one of the reasons I haven’t posted about it sooner). One has to decide whether all the characters are biglyphs or just some of them, and whether the others are ligatures or monoglyphs.

It’s also important to have some sense of whether the spaces are real or contrived and one has to figure out if the text has been abbreviated. If it has, could this group of four glyphs be the anchor around which the rest of the text is crafted?

Vowels aren’t really necessary for text to be comprehensible. Mst ppl cn fgr t txt wtht vwls and many languages were originally written without vowels. What else might cause four biglyphs to share certain commonalities in shape and behavior?

Can we find out more by looking at where they appear in the manuscript?

It may seem as though individual glyphs are more prevalent in certain sections, but keep in mind that the big-plants section is extensive and the amount of text on unillustrated starred-text pages is considerable, so it is natural that they would show up more often on these folios. However, it’s interesting to see the consistency with which the first four show up throughout the manuscript and how they differ in overall balance from the last two.

It may be noteworthy that ax occurs less frequently on the big-plant pages than the previous three and that EVA-dy, despite its relative frequency, is very infrequent on the rosettes foldout compared to the first four other five. I’m not even sure that EVA-dy is a biglyph. It might be a ligature plus a null.

I have much more information on the structure of the text but that’s probably enough for one blog. Once you begin to notice these pairs, they  jump off the page and you really can’t help wondering if Voynichese is synthetically constructed.

J.K. Petersen

© Copyright 2017 J.K. Petersen, All Rights Reserved

That Funny 4 Glyph

This article describes the odd glyph that resembles the number “4” (EVA-q). It’s odd because the “4” shape wasn’t prevalent in the late 1300s and early 1400s. It was a transitional period when many scribes were still using a character that looks like EVA-l to represent the number 4 and a few were still using Roman numerals.

That’s not to say that the VMS “4” is a numeral, I’m simply pointing out that this particular shape, with a sharp angled corner, wasn’t a common choice to represent a letter or a number when the VMS was created. Nevertheless, I believe it has its roots in Latin.

[Note: This is the more complete version of some images and commentary I posted on the forum a few months ago describing the glyph known as EVA-q. Even this is not the complete story, as there are statistics to go with the images, but it’s far too much information for one blog, so this article focuses on possible origins of the shape.}

How 4 Manifests in the VMS

The 4 glyph makes its first appearance on folio 1v, and from that point is frequently at the beginning of word-tokens and is followed by “o” about 90% of the time. The VMS is so regular in its construction, it would be tempting to think the other 10% are transcription errors, but the 4 has some interesting properties that suggest these are choices rather than errors. But first, here are some examples of the 4o combination, since it is most prevalent. Note that it is usually at the beginnings of Vwords.

But not always, a 4o can show up in the middle or at the end.

It’s often assumed that “4o” functions as a unit (and perhaps it does), but 4 is not always accompanied by “o”. The following examples show that 4 can be followed by other glyphs, such as a “c” shape, a “c” with a tail, a bench character, a benched gallows character, or the “cap” that represents missing letters in Latin. As further examples, the 4 can also be followed by a benched gallows (f103v) or an “i” (f106r), and 4o itself can be followed by “o”.

Both 4o and 4’o can stand alone—they don’t have to be attached to other Vwords. Note also, in the above and below examples, that a Latin abbreviation mark is sometimes associated with 4o. In the above example, the symbol is curved, but it is sometimes written as a straight macron-shape rather than a curved one, and occasionally there is a mysterious extra line connecting the two glyphs (might this be a hidden macron, or a combination with a different meaning?).

Whether two different abbreviation symbols have the same or different meanings depends on the scribe. Some were quite precise in the way they represented missing letters, others used whatever was convenient to the hand (or their imaginations).

The 4 is frequently followed by “o” (at the beginnings of Vwords), but they are not necessarily a combination—”o” sometimes precedes 4, or is sometimes combined with another 4o.

A “4” By Other Names

This glyph is often called “q” because some have interpreted “4o” as “qu” (it is also mapped to “q” on the EVA system but this keyboard position was not intended to impose meaning on the glyph). Sometimes the 4-shape has a soft connection rather than an angular one, making it look more like “q” than a “4”. Note that the pic on the right has a sharp-angled “4” on the same line as a soft “4”. Sometimes it’s indistinguishable from a “q” (assuming this is EVA-q and not EVA-y—sometimes it’s hard to tell).

There is more than one way to interpret the variation in the loop of the “4” glyph. Perhaps the soft-4 and the sharp-4 have different meanings, or perhaps they don’t, just as a “p” sometimes has a loop that connects and sometimes doesn’t, but means the same thing.

There is more than one character directly associated with 4. Sometimes the 4 is attached to a glyph that resembles the letter ell or the Latin “-is” abbreviation. This combination strongly resembles a mini-gallows character with a descender. The resemblance is so strong, you have to wonder if there’s a connection between EVA-q + “-is” and EVA-k, either in terms of glyph origins or meaning. Or is this a way to hide two consecutive gallows characters? It’s hard to test an uncommon combination—there aren’t enough instances to know if it behaves in the same way as 4o.

Common Patterns

When 4 is combined with o, it frequently precedes a gallows character and the gallows character frequently precedes a or c shapes. I’ve described this rule-like characteristic of Voynichese in past blogs.

Note that 4o is usually in front of the H-like gallows, not the P-like gallows. Note also that some of these are soft-4 and some sharp-4 and yet, at least superficially, they appear to behave in the same way.


If the VMS were Latin, then 4’o (4 and o with a straight or curved macron) can be interpreted in a number of ways—there’s no specific rule for how to expand the abbreviation symbol and there was quite a bit of variation in how scribes drew these squiggles, curves, and lines, but there were some general guidelines.

For example, a “squiggle” like the one found on the first page of the VMS is often interpreted as “er”, “ir”, “re” or “ri”, but even this symbol is sometimes used for other letters. Thus, in a medieval manuscript, one would look at neighboring words (in this case talis and est) to determine whether q’o represents “quero”, “quo”, “questo”, or “quomodo”.

You might also notice in this example of 15th-century cursive that the “q” shape isn’t round, it’s quite angular, almost like a VMS 4, but it was less frequently written this way.

Does This Mean the Voynich Manuscript is Latin?

Many have tried to translate it as such, it’s one of the most commonly claimed languages in VMS history, but most attempts range from shaky to bad, and sometimes they are really bad (I’ve only seen one that strikes me as a reasonable effort and that’s the one by Yulia May). So far, we only know that the glyphs are Latin, not that the language was Latin. Latin scribal conventions were common to many languages, including Greek, French, Spanish, German, Italian, English, Bohemian, and Scandinavian. The shapes by themselves do not reveal the language—they are adapted to represent common linguistic patterns in that language. Thus, a sign that means “-us” in Latin could potentially be used to represent “-en” or some other common ending in German.

In fact, we still use this system in English. The letter “w” with a line over it or a swooped-back tail is an abbreviation for “with” in the same way that an “a” with a swooped-back tail represents “aut” or “autem” in Latin.

So Where Did the 4 Shape Originate?

It’s possible that the VMS 4 is simply an invention, that no particular precedent inspired the shape. Or maybe the idea came from noticing quirks in the handwriting of certain scribes. As an example of how the “p” was sometimes written in medieval times, notice how the loop in this example is almost completely disconnected from the stem—it almost looks like 4o.

It’s tempting to think this might have twigged the idea for the VMS 4 glyph but, based on the way the abbreviations symbols are associated with 4o, I suspect the true inspiration might be another Latin abbreviation.

The 4 Glyph With and Without Ascenders

Note how the following VMS glyph resembles a 4 and appears to behave as a 4 when it precedes an o, but has an extra-long ascender-like stem. The VMS scribes were clearly familiar with Latin scribal conventions, but one still needs to consider whether this is scribal habit, a purely physical error, or a letter that started out as a gallows and got changed to 4. If it is a slip from Voynichese to regular Latin, does it reveal something about the glyph?

Unfortunately, there aren’t enough instances of the ascender4 to know, but we can take a look at other scripts to see if the shape was extant.

Historical Precedents

The VMS ascender4 reminded me of a sample of Visigothic text that includes a number of Latin abbreviations.

  • First note the macron in the shape of an old-style four (it looks like an x with a loop on top) near the end of line seven. It’s basically the same shape as EVA-l.
  • There is also a q with a long s-curve crossing the stem on the ninth line that can stand for various words including “quo”. A similar convention when applied to a p can turn it into “pre” or “pro” depending on whether the line is straight or curved.
  • There is a shape on the bottom that resembles a backwards gallows P that has various meanings depending on the time period. It can mean -us or -rum and is sometimes similar to a pilcrow except that it marks the end of a paragraph rather than the beginning.
  • There is an ampersand on line six near the beginning that can stand for “et” (as in Latin “and”) or for the two letters e and t if used as a ligature.

These are all common abbreviations. But the one of particular interest, circled in red, is one that matches the shape of the VMS ascender4. It can be attached to many different letters and is usually at the ends of words.

This character is comprised of a c-shape that loops over a long vertical stem. The loop is sometimes sharp, like a 4, or soft, like a q. The sharpness of the loop does not change the meaning of the symbol. Here it is primarily attached to “q” or “l” but it can be used in many different ways.

Typically the shape represents “-us” (which eventually evolved into a “9” shape or an apostrophe in later medieval manuscripts), but it can stand for other common endings that can be discerned by context, including “-uibus”. If it has a small extra loop on the top right, it can also mean “per” (which was later written by placing a line through the stem of a p rather than extending the top).

It’s possible this abbreviation inspired the shape for the VMS ascender4 and possibly also the 4.

Assuming there is meaning behind the VMS text, this symbol could potentially be expanded into a variety of letter patterns. In Latin it typically represents an ending, but it could just as easily be used as a prefix. Or, alternately, perhaps it does represent an ending and Voynichese is read right to left, even though it has been written left to right. Whenever I examine the text, I always try to scan it in both directions and not make too many assumptions about direction.


Taken individually, it would be difficult to determine the exact origin of a glyph, but when the VMS characters are studied as a whole, a strong pattern of Latin letters and abbreviations emerges. I haven’t had time to write up all the glyphs yet, I’m adding them as I can make time, but I have found abbreviation origins for almost all of the more peculiar-looking glyphs—and they trace back to Greco-Roman scribal conventions.

I don’t know if the ascender4 is based on the abbreviation-glyph illustrated above (or even if ascender4 and 4 are related), but it might be, so I thought it worth providing an example. If it is, then there’s still the challenge of figuring out whether the shape is simply a shape, a character (alpha or numeric), or something intended to be expanded into additional letters.


J.K. Petersen

© Copyright 2017 J.K. Petersen, All Rights Reserved




Sense or Non-Sense?

In Nick Pelling’s Cipher Mysteries blog, he commented on the challenges of parsing VMS text and creating transcripts, and specifically noted:

“… a big problem with entropy studies (and indeed with statistical studies in general) is that they tend to over-report the exceptions to the rule: for something like qo, it is easy to look at the instances of qa and conclude that these are ‘obviously’ strongly-meaningful alternatives to the linguistically-conventional qo. But from the strongly-structured point of view, they look well-nigh indistinguishable from copying errors. How can we test these two ideas?”

This is indeed one of the challenges in transcribing and understanding Voynichese. Our perception of the structure of the text will be skewed unless one can sort out, to a reasonable extent 1) the exceptions/rare forms, 2) handwriting variations, and 3) copying errors, from what may be meaningful text, so that relevant variations are acknowledged and artifacts filtered out.

The Characteristics of EVA-qo

The subject of EVA-qo was touched on in my previous blog, in which I posted a variant 4o image that shows a possible “component” relationship between “qo” and glyphs with ascenders. Prior to that I expressed uncertainty about identifying when EVA-qo functions on its own and when it functions as a pair (I suspect that pairs and singles may function according to priorities), but more examples are necessary to cover the topic in depth.

Glancing through the VMS, one will notice that “4o” is a frequent combination. In the following clip, which I chose arbitrarily, one sees several examples of 4o within the space of a few lines. One stands alone (which happens more often than one might think), the others are at the beginnings of V-words. Notice how some have sharp points and others are rounded. Most of them connect to the following glyph:

How does one determine if the 4 and o are intended as a paired glyph, or whether it is simply a common combination such as “qu” in English? Do the sharp and rounded corners have any significance? or the connected/disconnected characters? Note how 4o is frequently followed by an ascender glyph, except for EVA-qol. EVA-ol is one of the combinations that may function as a pair, in which case one has to ask whether 4 can function as a “single” when followed by a pair, according to some rule of precedence, as was noted in the discussion of pair patterns.


At first glance, it might appear that 4 is always followed by “o” and always falls at the beginning of a word. In fact, 4o can occur at the ends of words and occasionally in the middle.

Many characters can follow the 4, including a common Latin abbreviation symbol (which is sometimes straight, sometimes curved). Here are some examples:

It’s also fairly common for 4 to be preceded by o or 4o, and 4o4 and o4o sometimes stand alone:

The o4o words appear mainly in the plant, pool, and starred-text pages, with one in cosmology and one on map rosette #1. There are none in the zodiac or small-plant pages.

Some variations differ much more than those with straight or rounded connections, as in this example that I’m reposting from the previous blog. It has an extended stem and, below it, a variant that is followed by an “l” shape rather than “o” such that the glyph bears a strong resemblance to a 1.5-legged ascender:

To show this in context, note how a shift in position determines whether this combination looks more like 4o or a 1.5-legged ascender glyph. This isn’t drawn like a malformed 4o or oddball gallows glyph, this looks deliberate, but notice how it falls immediately before ascender glyphs or one that is a common pair, a position typical for 4o:

When EVA-q is followed by a form that looks like a cursive ell, it resembles a 1.5-leg double-looped ascender, except that it is positioned as EVA-q would be, as descending below the baseline.

The 4 glyph doesn’t only resemble the left leg and loop of an ascender, sometimes it is difficult to distinguish a rounded form of 4 from a straight-leg form of EVA-y, both of which look like a Latin “q”.


And Now to the Numbers

The 4 glyph makes its first appearance on folio 1v (the second page, as the VMS is currently bound), paired with “o”, with a line above it. If this were Latin, the line would indicate missing letters in much the same way as we use an apostrophe.

On folio 2r, “4o” becomes more numerous and precedes a variety of glyphs, with ascenders being the most common.

On folio 5r, something interesting happens. There is a unique word on the 6th line (EVA-qokeeey), but if you remove the 4o, it appears as a unique word, without the 4, on another large-plant page (folio 49r) and, without the 4o, on plant page 50v. Similarly, unique word qoToldaiin (folio 4v), without the 4o, appears as a unique word on folio 67r1.

It’s been suggested that unique words are names, but if they were names, wouldn’t someone have decoded them by now? And would so many names, differing only in the first one or two characters, appear on seemingly unrelated pages? If they are names, such as names of plants, wouldn’t they show up elsewhere in the manuscript, rather than being unique? It’s typical of medieval manuscripts to be extremely repetitive, especially if they include recipes, charms, or classification systems—the same names appear with great frequency, especially if they are common ingredients.

I haven’t seen any successful attempts to resolve unique tokens into natural language in any consistent or generalizable way, so maybe they aren’t words. Perhaps they serve a nonlinguistic function. Assuming the spaces can be believed, and they are indeed unique, is it possible that a certain class of word-tokens represents a medieval rendition of pointers, patterns that relate one data location to another?


The “4o” words are not all unique, some are quite common. For example, qokaiin occurs more than 300 times, mostly on the plant, pool, and starred text pages—it does not appear on the zodiac or rosette pages, which argues against random generation of the text. The 4o words tend to appear only once on the zodiac pages, except for Gemini and Sagittarius, where they occur several times. A unique word on the Pisces page (qoTeeal) appears as a unique word without the q on 69v, a cosmology page.


If the VMS includes a network of relationships, then it’s essential to determine if the glyph variations are meaningful and whether the spaces are real or contrived. As an example, is the unique word qoToldaiin, on plant 4v, related to the component words qoTol,daiin  that appear next to each other on folios 19v, 21v? The first one has a sharp-4 and an ambiguous space. The latter two, have sharp-4 and very clear spaces.

I have much more information on individual glyphs, but this is more than enough for one blog. I’d like to close with a suggestion that “confidence levels” for certain variations be documented in some way (for example, a pointed or rounded q might not be significant, but q with a high ascender is sufficiently different that it might), and a strong suggestion for structuring VMS transcripts to include Quire X, Side X, Folio X in the explanatory sections for each folio. That way, when looking at glyph variations and V-word relationships, it’s easier to see if similarities and differences are tied to physical proximity.

J.K. Petersen

© Copyright 2017 J.K. Petersen, All Rights Reserved

On the Gallows

VMS characters with ascenders have some interesting properties beyond their frequent appearance at the beginnings of paragraphs. Unfortunately, the seemingly simple task of selecting the most representative shapes and slotting them into a grid took far longer than I expected and this blog languished in “draft” status for several years. I did manage to post a preliminary assessment of whether some of the glyphs might be pilcrows, but did not go into depth about the glyphs themselves because I wanted to discuss that in a separate blog as follows…

Interpreting the Text

My perception of VMS glyphs differs substantially from one of the more popular transcripts created by Takeshi Takahashi. So much so, that I ended up creating my own rather than using any of the ones that were extant. I wanted to correct a number of errors, add the labels, and allow for the possibility that certain variations might be meaningful rather than the result of scribal variations. What follows is not just a collection of glyphs with differing shapes—it’s the result of a long process of trying to statistically differentiate variations that are meaningful from those that are not.

The “Gallows” Characters

As mentioned in previous blogs, most of the VMS glyphs can be traced to Latin characters and abbreviations, with a few that are derived from Greek, but there are some tall glyphs that are sufficiently different to modern eyes that they have been dubbed “gallows characters”. Not everyone is happy about the “gallows” moniker, so I’ll mostly refer to them as VMS glyphs with ascenders. Here is a sample of some common shapes. Note that some have one loop and some have two:

The VMS ascenders are not entirely alien. If they are ligatures (more than one character joined to create a cohesive shape), then the two-legged variety (right) resembles the Latin abbreviation for “-tis” or “Item”. The “P” shape could come from anywhere. A “P” shape is common to many alphabets, including Greek, Latin, Armenian, Cyrillic, and others, but I suspect that a rational design process may account more fully for the VMS “P” than any particular alphabet, an idea that I’ll describe below.

The Takahashi transcription recognizes four basic kinds of ascenders, the single-leg glyph (left), the double-leg glyph (right) and the double-looped version of each. Added to this are “benched” versions of each of the shapes—those that have a crossbar near the base that connects glyphs on either side. These are designated in the transcript with an uppercase “Z” that follows the character for the basic form, which is not a bad way to do it, since it allows for searching both the benched and unbenched varieties, and uses a shape that is easy to remember as the rotated Z somewhat resembles a crossbar.

Others have noticed that the crossbar does not always connect on both sides and have tried to account for this in various revisions of the EVA fonts.

Discerning Intent

I would like to propose that this appealingly simple classification scheme utilized in the Takahashi transcript may be wrong. When you really study them, there’s more to the ascenders than meets the eye and pen variations may, in fact, be meaningful.

Those who are familiar with Hebrew and some of the Malaysian scripts, already understand that subtle differences between characters might change their meaning. The swoop of a tail, the presence of a dot or tick mark that is high, medium or low, or the length of a crossbar, represent different letters or syllables. I think some of these variations may also exist in the VMS characters, but one cannot judge solely on shape, one has to look at the distribution and position of the glyphs, and the way they are written by different scribes.

Before discussing this in depth, I’d like to point out some morphological similarities between the two more distinctive ascenders. Note how the two-legged ascender resembles a one-legged ascender whose drawing was interrupted before the second leg was finished:

If you are skeptical that this may be the basis for the “P” shape, consider examples 5 and 6 below, in which the “interrupted” leg is clearly visible, a glyph that is neither a single- or double-leg, but one that is in between:

In the images that are second and third from the right, note how the second leg does not always reach the baseline or swoop back in a tail. The leg second-right is attached to a c-shape, a combination that occurs less ambiguously in other parts of the manuscript. This suggests the shapes might be ligatures, rather than individual characters and that the 1.5-legged glyph may be a component in its own right.


There’s more to this 1.5-leg idea that might surprise you…

Take a look at these 4o glyphs (EVA-qo) that have a long stem that rises up above the leading c-shape (not all 4o glyphs have a rounded head, some are very sharp, but these are all round). Then look at the shapes below them that are constructed the same way but are followed by an “l” shape rather than “o”.

Did you notice how this “4o” variant on the second line below resembles a 1.5-leg gallows? The main difference is that the shape on the left is a descender rather than an ascender. If it were shifted upwards, it would be interpreted as a 1.5-leg gallows without an additional letter or crossbar, but it superficially resembles 4o because it appears at the beginning of V-words and is mostly below the baseline. The long-stemmed “4” may be distinct from other glyphs identified as EVA-q:

I don’t believe the various theories that there are microscopic encodings in the VMS, based on incredibly tiny variations in shapes—I’ve seen no convincing evidence that this is so. I do however, believe that some of the more overt differences that might be meaningful have been overlooked and might make the text look more repetitive than it actually is.

The Case of the Curly Tail

Something I’ve wondered since I first examined the VMS is whether the curled tails on ascenders and other characters are meaningful. In Latin, a curled tail on the letters “i” or “r” are significant. They indicate abbreviations such as “er/ir/re/re” and sometimes “us/um”. On a “P”, they differentiate between “pro” and “per”.

Some scribes even control the shape of the tail to differentiate between “er” and “ir”. You’ll notice in the self-similarity example above that the far-right glyph has a distinctly curled tail. The one in the picture below is clearly straight. There are some that are slightly ambiguous but most of them can be differentiated as one or the other.

Does the difference matter in the VMS? I suspect it does. Curly tails appear more often on ascenders that are in key positions, such as the beginnings of paragraphs or places where one might expect a capitulum, but it’s difficult to know from position alone whether this is their function. Most of the midline ascenders do not have curled tails. This is not the way “per” and “pro” behave in Latin. They are liberally sprinkled throughout the text, with “per” being a bit more common than “pro”, but not excessively so, so it appears that the VMS glyphs are similar to Latin in shape but not necessarily in function.

Notes About the Chart

What follows is a PDF file with one possible configuration for studying and classifying VMS characters with ascenders, based on variations that may distinguish one from another.

  • As far as I can tell, the length of the tail is not significant, except possibly when it touches the baseline on a standalone glyph, or the bottom of the stem on a combination-glyph.
  • The shape of the tail (straight or curved) looks like it might be significant.
  • There appears to be some consistency in the way the connections are made on one-legged ascenders. The top-left connection might be significant. The bottom right corner, where the tail is attached, is sometimes sharp and sometimes rounded, but this connection doesn’t appear to be significant, as far as I can tell.
  • The characters that connect to the crossbars on “benched” characters are quite variable and the variation appears deliberate. The straight benches do not appear to be corrupted c-shapes. They might represent EVA-i, or EVA-r without the tail.
  • The combination glyphs appear to be a gallows character combined with lowercase glyphs or two gallows glyphs combined, perhaps as a way to put two together without setting them next to each other. It’s my feeling that combination characters differ from embellished ascenders, but I am not certain yet.

This is a preliminary chart, subject to revision. You can click on the link under the thumbnail to download the PDF file:


(New chart uploaded June 3, 2017, with one error corrected. June 4, 2017, Gap bar moved next to Touch bars.)

                                                                                                                                   J.K. Petersen

© Copyright 2017 J.K. Petersen, All Rights Reserved

Two of This or One of That?

One of the difficulties in creating a transcript and analyzing textual patterns in the Voynich Manuscript is the ambiguity in some of the characters. When this occurs in common words, it makes it more difficult to assess glyph relationships and frequencies.

A simple example might illustrate this problem. I mentioned in my previous blog that I believe the paired c-shapes are meant to be read as one character (in most instances). There is also a single c (EVA-e) which may occur next to a double-c, to create three in a row. When there are three in a row, how does one decide whether it’s three cees, a double-c following to a single-c, or a single-c following a double-c?

In this example, I’m leaning toward the VMS glyphs (top) being two double-c shapes because of the slightly larger gap between the two pairs and the way the cc behaves in other parts of the manuscript, but I’m not 100% sure because the two latter cees are more tightly written than the first two. Is this normal pen-variation or are the first two cees single cees followed by a double-c?

Sometimes all we have to go on is slight differences in the spaces between characters and that’s not a good way to do it—there will always be some uncertainty, which is one of the reasons I feel it’s important to study the rule set and possible pairing paradigm for the VMS. Then the context can help us determine which glyphs are intended as ligatures and which might function as pairs.

The Devil in the Details

Unfortunately, ambiguity exists in one of the most common VMS word-tokens, one that is popularly called “dain”. I don’t use the EVA font-set, I developed my own based on shape designations, but you should be able to see the correspondence in the following illustration fairly readily.

In this example, there is ambiguity in the straight shape that alternately resembles a double-i or possibly a “u” as it was often written slightly separated, with straight legs, in the middle ages. I use a “v” and sometimes a “w” to describe the ending shape with a tail but I make no assumptions about what these shapes mean or whether the swept-up tail indicates an abbreviation, as it would in classical Latin, or whether it is an embellished glyph designed to look like Latin, just as the “9” shape (EVA-y) morphologically and positionally follows Latin conventions:

To complicate matters further, there are places in the manuscript where there is an additional stroke between the a-shape and the swept-up tail, one that Takahashi (and perhaps other transcribers) sometimes missed.


A fresh transcript is needed, and not just a “corrected” transcript that makes better assessments of the spaces (I’ve noticed errors in which glyphs with clear spaces around them have been attached to nearby words), but one in which all the glyphs are included, even ones that “look funny” because there are so many in a row, along with consideration for alternate interpretations for ligatures (combined glyphs) and paired glyphs.

I created a transcript that corrects some of these problems, but it’s not a stand-alone file. I’ve integrated it with a set of self-made VMS fonts and applications so that the whole thing is an interdependent set of tools that can’t really be split apart as they currently stand.

I can make some suggestions, however. When I created my fonts, I put the VMS characters in the upper register and the regular characters in the lower register as they are usually typed on the keyboard, so that it’s seamless to combine VMS and regular characters in the same document (handy if you’re writing an article about the VMS). This also allows comments to be added to the transcript that don’t interfere with searches of the VMS glyphs. Unicode standards have plenty of space for this, and it’s not difficult to come up with mnemonic references to the shapes to make it easier to type. I also set up glyphs that are similar such that they can be searched together or separately. Adding a symbol to the glyph is usually a better overall solution than putting each variation of a basic form in a different font-slot, a point that I’ll discuss more fully in my next blog.


                                                                                                                                   J.K. Petersen

© Copyright 2017 J.K. Petersen, All Rights Reserved

Construction of the Voynich Manuscript Text

In previous blogs, I gave quick examples of the rule-dependent way in which Voynich Manuscript glyphs are combined (it’s far beyond the scope of a blog to define the entire rule-set, so I used a single page of text as an example and have been working on a long paper that describes the overall manuscript more fully). I also pointed out some of the more common atomic units, as I think of them.

Since that time I’ve been trying to think of a way to make these patterns and relationships easier to understand.

Hopefully this visualization method can illustrate why computational and linguistic attacks that assess individual glyphs may not yield fruitful results. The VMS has very particular ways of combining glyphs that affect not only which ones appear next to one another with greater frequency, and in what order, but also controls word-length in unique ways.

  • Note, as mentioned in the diagram below, a VMS double-c-shape (adjacent EVA-e glyphs) appears to function as a single unit in much the same way as a double-c shape in Carolingian script (right) represents the single letter “a”.
  • Certain glyphs, like EVA-d and EVA-s appear to function as single glyphs unless paired in very specific ways (with certain glyphs in certain positions, such as EVA-dy at the ends of words).
  • Note the prevalence of combinations like EVA- or, ar, oI, 4o, ly, che, sho, and ey. These pair-syllables, in various combinations with glyphs that function as singles, characterize the entire manuscript, including text in labels and wheels.
  • Note also the difficulty of assessing whether 4o is intended as separate glyphs or as a pair (in some tokens, it could be assigned either way and there are a few other combinations with this characteristic). I have included examples of both possible interpretations of 4o-tokens in the following illustration with the caveat that I am less sure of the 4o- breakdowns than most of the others.

Despite the difficulty of distinguishing singles from pairs with complete accuracy, I think these short examples come close and may help illustrate how the VMS text differs from common natural language patterns and patterns evident in medieval ciphered texts, and especially why one-to-one substitution systems have so far been unsuccessful.


Also, give some thought as to how paired glyphs affect entropy and word length…

Paired glyphs greatly increase the number of letters or sounds a system could potentially represent. For example, if you had only o, a, r, and x, and placed them in a grid as pairs, your four glyphs could yield 16 pair-glyphs plus the four original glyphs to represent 20 letters or sounds. There aren’t as many combinations as this in the VMS, because glyph order is deliberately restricted and it’s not practical to put mirror pairs next to each other as they are hard to distinguish without extra spaces, but even so, the concept applies—entropy increases

As to word length… the VMS word-tokens are already short compared to natural languages, but if some of the glyphs are paired, word-length decreases further. If one is looking for letter or sound correspondence in text that has a large number of paired glyphs, then it’s more likely that they represent syllables, fragments, or abbreviations, rather than full words.

So, enough discussion… here are two examples that I grabbed arbitrarily. They’re short, but hopefully long enough to get the ideas across.

You can click on the image to see it full-sized  (you may have to click again when the new tab opens to read the small print):


Postscript (after getting some much-needed sleep): I hope it is apparent from my previous comments that these are examples, not a definitive breakdown. In the illustration, I have broken down the “4o” words and some of the “9” words (EVA-y) in both ways to show both possibilities—with the 4 and 9 as singles and as pairs, because there is evidence elsewhere in the manuscript that both are possible interpretations. Some pairs (the common ones) are much more consistent and discernible than 4 and 9 word-tokens and I have a long list of stats for some of the more consistent pairs.

The distinction is important because pairs and singles may have different classes of meaning. For example, in Latin, the 9 character (which frequently functions as a single in the VMS, except when paired with EVA-d and possibly EVA-e) expands into prefixes like con- and com- and suffixes like -us and -um, which brings up the question of whether pairs might represent letters and singles might represent abbreviations, as were commonly used in medieval scripts, or (another possibility) whether they were intended to be differentiated in some other way, such as singles representing nulls, modifiers, or markers and pairs representing something else (letters, sounds, or concepts). Note that the singles are often at the beginnings of paragraphs and word-tokens, and also sometimes form one-glyph word-tokens. It is further possible that the high preponderance of “o” glyphs (particularly those in the first position) might be evidence of a pairing process intended to make tokens come out in a certain way (with a particular pattern or length).

All this assumes, of course, that the VMS text is meaningful, something that has not been proven. The pattern of pairs and singles could just as easily have been devised to make it easier to write meaningless text that looks like syllables and abbreviations. I still have a certain cautious optimism that there is meaning behind the text and will post another blog soon that explores some of the details of gallows characters that haven’t yet been discussed.

                                                                                                                                   J.K. Petersen

© Copyright 2017 J.K. Petersen, All Rights Reserved