Mixed-Up Musa

Did medieval illustrators confuse Musa (which we now know as banana) with another plant labeled “muse”? Early drawings of “muse” fruits don’t look like bananas. They are small, with pointy ends. In fact, the taste of “muse” was likened to cucumber or melon, fruits that are much more watery than banana. Yet as time went on, the descriptions, drawings, and the name were changed to “Musa”, which we know as banana/plantain plants.

Bananas and other eastern or African semitropical and tropical plants were probably known to Europeans through word-of-mouth, just as they had heard of hyenas but apparently had not seen them. Medieval drawings of hyenas are quite fanciful, and sometimes understood only by their labels. Is this what happened to banana plants? Did they invent the drawings and call them “muse”, or are the early herbals depicting some other plant that became confused with banana plants as knowledge of the banana gradually spread through Europe?

Sugar cane was known in Europe from early times (although it was not imported into northern Europe until later), probably because it’s easier to ship than many kinds of plants, but bananas are not easy to ship. They bruise and ferment and thus were not as familiar to Europeans as sugar cane in the early Middle Ages.

Historical Background

There were a number of plants locally called “muse/musse/mus” in the Middle Ages—field plants like “mouse ear” that are similar to Hieracium, but they don’t look like the 14th-century drawings labeled “muse/musse” in manuscripts such as the Manfredus herbal, or in the various copies of Tacuinum Sanitatis.

So if “muse” is not the mouse-ear plant (which is often included in herbal manuscripts under other names, and usually drawn fairly accurately), and it’s not banana, then what is it?

Lining up the Illustrations

Maybe Egerton 747, the Manfredus du Monte herbal, and versions of Tacuinum Sanitatis can provide clues. There is a plant in the Manfredus manuscript that has an analog in both form and name that might easily be confused with banana.

First note how “muse/musse” is drawn. It’s not a good representation of a banana plant. It has a basal whorl of upright lanceolate leaves, similar in shape to Inula, and nut-like dangling fruits sometimes painted red, sometimes yellow (this color difference is significant, as will be illustrated later). Banana plants might look vaguely like this when they first start to sprout, but the stalks grow substantially before they bear fruit and the fruits do not hang on long stalks with skinny petioles above the leaves.

Sometimes the fruits of “muse” are drawn a little more capsule-shaped, with striations, but these too are quite different from banana. The image on the left is from the Manfredus herbal (late 14th century), those in the middle are from two different versions of Tacuinum Sanitatus:

In the text that accompanies earlier drawings of “muse”, the leaves have been compared to Inula (which does not resemble banana leaves), and the taste has been compared to melon and to citric fruits (which doesn’t match well with banana either).

So there are several problems with assuming that “muse” is Musa (banana):

  • The name is different.
  • The leaves are shaped and positioned differently.
  • The fruit stalks are much too tall and the fruits too discrete and the wrong shape for bananas.
  • The description of the taste doesn’t fit banana.

In later herbals, the descriptions start to sound more like banana (based on the assumption that “muse” was an alternate spelling for “musa”) and the name is changed to “musa”. Eventually drawings and descriptions of real banana plants are substituted.

Other Possibilities

Is there a plant that matches “muse” better than banana?

Yes. I don’t know if it’s the correct plant, but melegueta matches better to the drawing and description of “muse” than the banana.

And here’s the interesting part (and another reason it might be confused with banana). Banana became known as Musa paradisa. The melegueta plant is known as “pomum paradisi” (apple of paradise). Thus, if someone unfamiliar with melegueta saw the name “muse” in combination with “paradisi” they might assume it was a banana plant.

Here are some images of melegueta and cardamom. Melegueta is a west-African plant with lanceolate leaves. It has slightly dangling fruits on spindly stalks, and the ripening fruits are often red or yellow:

pic of melegueta plant

Comparison of “muse” plant with leaves and fruits of pomum paradisi, the melegueta plant. Image credits: Bibliothèque Nationale de France, gernot-katzers-spice-pages.com, tropical.theferns.info, sciencedirect.com.

Medieval illustrators aren’t the only ones who appear to have confused “musa” (banana) with early depictions of a plant called “muse”—15th-century annotators and 17th- and 18th-century commentators made the same mistake. No one appears to have critically evaluated the way medieval descriptions and drawings changed over the course of two centuries from “muse” (pomum paradisi) to Musa paradisi (banana), nor have they investigated what exactly was meant by “muse” in the earlier herbals. It has long been assumed that the old herbals, even ones with good illustrations, specifically included erroneous descriptions and bad drawings of banana plants rather than reasonable drawings of a different plant.

The drawing and description of “muse” is similar in shape, proportion, name, and colors to melegueta and a few other Amomum species, but there’s more…

I found two additional pieces of evidence to support this tentative ID.

  • The first has to do with taste, which is described in one early medieval source as “citric” (I don’t think anyone would describe the taste of banana as citric). The Wikipedia entry for melegueta describes the taste as peppery “with hints of citrus”.
  • The second is a sequence of plants in Egerton 747 (c. late 13th century), a reference that bears many similarities to the Manfredus herbal (some say Manfredus is copied from Egerton 747, others say both are copied from a common source). In Egerton 747, is a series of shrubby trees, beginning with Nux muscata and ending with Nux vomica. Note that it includes Nux sciarca (melegueta):

In the Manfredus herbal, the mystery plant “muse” follows “muscata”. If “muse” is meant to represent melegueta, it’s more accurate than the picture of melegueta in Egerton 747, but what is even more significant is that it follows closely to muscata. Manfredus does not correspond 1-to-1 to Egerton plants, but the contents and sequence are very similar.

Melegueta (Aframomum grana-paradisi) is a medicinal plant in the ginger family with anti-inflammatory properties. It is native to coastal west Africa, and is related to cardamom. Its inclusion in Egerton 747 suggests that it was known in Europe at least since the late 13th century (possibly through Arab traders).

There are several Amomum plants that should be considered together with melegueta, including Amomum villosum, Amomum cardamom, and Amomum compactum (false cardamom).

Note that the Manfredus and Tacuinum drawings of “muse” resemble Amomum plants more closely when they are younger. Melegueta leaves tend to become more palm-like as the plant grows, although some of the other Amomum species remain more lanceolate, like the drawings above.


I’m not sure how a name like Nux melegueta or Amomum might become “muse”*, but it’s not surprising that “muse” (pomum paradisi) might be confused with “musa paradisi” (banana).

Name changes are not uncommon. “Earth apple” used to refer to Cyclamen, a common medicinal plant with a big lump of a root. After the conquest of America and import of New World plants into Europe, the potato came to be known as the “earth apple” (pomme de terre) and Cyclamen gradually lost the name.

But I don’t think this is a case of the name changing. I think it’s the result of confusion…

Aframomum and Musa (banana) plants were not well-known in the west and had similar names (pomum paradisi vs. musa paradisi).

The 13th- and 14th-century drawings and descriptions of “muse” match well to melegueta and a few of the other Amomum species, and less well to banana. By the middle of the 15th century, traditional descriptions were altered to better fit the banana and the name was changed to “musa” and, finally, by the 16th century, the name “musa” was paired with drawings of actual banana plants and the name “muse” seems to have disappeared.

J.K. Petersen

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

Postscript, Nov. 10, 2018: One possible origin of the plant name “muse” that I forgot to include is hinted at in this quotation from classical Greek, which connects the Muses with a plant called “amomum” as follows:

“… Then he wove in Damagetus (a dark violet); Callimachus’ myrtle—sweet, but ever full of sour honey—; Euphorion’s rose campion; and the Muses’ amomum, who takes his name from the Dioscuri.”

It’s not enough information to identify the plant (Fée suggested Amomum racemosum, but there is no consensus of opinion) or to know if a connection between the muses and something called “amomum” still existed in the Middle Ages, but it might be worth filing for future study.

*Perhaps the answer is closer to the home of the plant itself. The Timneh tribes of Sierra Leone generally refer to the amoma/melegueta plants as “massa” (The Pharmaceutical Journal and Transactions, Volume 16, J. & A. Churchill., 1857. See also detailed discussion in Elements of Materia Medica).

The Ligature Legacy

“…some symbols in the Voynich Codex show similarities to letters found in sixteenth century codices from New Spain (Tucker and Talbert 2013; Comegys 2013) particularly the Codex Osuna (Valderrama 1600; Chávez Orozco 1947).”  —Janick and Tucker, Aug. 2018

The authors are talking about shapes that roughly resemble EVA-k and EVA-t. The following statement is much more surprising:

“We thus conclude that the author of the Voynich Codex made up his syllabary/alphabet, and the letters were borrowed from contemporary post-Conquest MesoAmerican manuscripts such as the Codex Osuma.”

In scholarly circles, “conclude” is a strong word—a word that needs to be backed up with solid evidence. Unfortunately, I find this conclusion highly questionable. The examples the authors use in their arguments are conventions that originated in Old World Latin scripts long before the 16th century. How can one use Old World scribal conventions to argue for a New World conclusion?

Is There a Preponderance of Evidence to Support the Conclusion?

Perhaps the authors felt that if the glyph shapes are taken together with botanical and biological identifications, there is enough evidence to support a New World origin, but the botanical and biological identifications of Tucker, Talbert, Janick, and Flaherty are highly questionable, as well. If you haven’t been following this discussion, then at least scan-read the previous blogs:

Even though the VMS 93r “sunflower” has a number of possible identifications (both New and Old World), Janick bases broad conclusions on this unproven ID as though it were fact:

“Simply put, there is no way a manuscript written on vellum that contains a sunflower and an armadillo could have been written before 1492,” —quoted on Purdue The Exponent news site, 10 Sep. 2018

There isn’t any proof of the identity of the “armadillo” either. It looks more like an Old World pangolin than a New World armadillo, but even this identification can be contested.

What I see in the papers and book by these authors is a collection of inadequately researched suppositions combined in a circular argument to support a New World theory. They pick out a few similarities and ignore the larger body of contrary evidence. They identify two completely different fish as the same fish. One of the plants they identified doesn’t even grow in MesoAmerica. They ignore numerous significant details like the cloudband under the “armadillo”. They ignore alternate IDs for the sunflower.

Unfortunately, the authors’ identification of Voynich-like glyphs suffers the same lack of critical evaluation as the plant and animal IDs, so let’s take a closer look at those.

The VMS-like Letters in the Codex Osuna

Here are examples of the Voynich-like glyphs cited by Janick and Tucker (and by Tucker and Talbert in a previous publication) in the Codex Osuna.

EVA-k is at the top, and EVA-t is at the bottom. Note that the handwriting is different:

Before you say, “Oh, those are similar”, make sure you read the rest of this blog. Bats and owls might look similar to a visitor from another planet, but one is a mammal, the other is a bird, and they are not closely related.

Visual Similarity is Not Enough (especially when they’re not actually that similar)

Something important Janick and Tucker did not mention is that the letters that appear to resemble EVA-k and EVA-t exist in two different scripts in two different languages. Failing to mention this distinction obscures the origin of these shapes, so I will fill in the missing pieces:

  • The EVA-k shape is in the sections written in Nahuatl.
  • The EVA-t shape is in the sections written in Spanish.

There are simple reasons for this, but they are important ones because there is no specific relationship between the Spanish and Nahuatl shapes. The similarities are coincidental, but some background might be necessary to make this clear…

Nahuatl Version of EVA-k

If you’ve heard the Bushmen click language, you know it can be very difficult to express this with Latin letters.

Similarly, there is a sound in Nahuatl that is hard to write. It’s made with the tongue against the back of the teeth, so the Spanish missionaries chose to represent the sound as the letters t + l and they wrote it as a ligature tl, with the crossbar of the “t” connecting to the loop of the “l”.

This ligature is not specific to Nahuatl or to the New World. It exists in Old World words like “atlas”, “battle”, “gatling”, etc. Note that the crossbar in the first letter “t” always extends some distance to the left of the stem, which is different from the way EVA-k is written:

It’s possible EVA-k is a ligature (two shapes combined) but if it is, then it follows age-old scribal conventions that are not specific to the New World (or to Nahuatl script). It doesn’t seem likely that VMS EVA-k was copied directly from Nahuatl if one goes by shape alone. It is more similar to some of the European ligatures and abbreviations such as “Il” (French) or “Item” (Italian, German, Latin) than the ligature on the left.

What About EVA-t?

Another common ligature in Old World languages that used Latin characters was the d” + “e” and since the letter “d” was written a dozen different ways, the “de” ligature is quite variable. A similar ligature combines “d” and “l” as in words like “headless”. Sometimes they are hard to tell apart from each other and from ligatures like “il”, but the concept is the same—two letters are combined so they can be written faster or in less space.

Here are examples of how “de” and related ligatures were sometimes written in Spanish scripts from the 14th to 16th centuries. The two on the right are from the Codex Osuna. The faster and loopier the writing, the more it resembles EVA-t (sort of):

The examples on the right illustrate how loose a ligature could be and how combinations like “de” or “dl” or “Il” or “Ie” need to be seen in context to be distinguished from one another, especially if it is an open-loop “d” followed by a very round “e” or “l”.

It has been suggested by Janick and Tucker that the glyphs above-right inspired EVA-t in the VMS, but this seems unlikely. EVA-t has long straight stems:

It’s possible the VMS char is a ligature, but even if it was inspired by “de” (I highly doubt that “de” was the inspiration but let’s pretend for a moment that it was), this ligature was common in many Old World languages.

The authors of Unraveling the Voynich Codex didn’t mention that the two shapes that resemble VMS glyphs are taken from two different sets of scripts (one in Nahuatl, the other in Spanish) and, more importantly, that these shapes were part of the normal scribal repertoire of Old World Europe and thus might have been seen by the creator of the Voynich Manuscript long before the conquest of MesoAmerica.


The authors didn’t provide any solid evidence that the inspiration for these shapes was specifically New World sources. In fact, the position of EVA-k and EVA-t within VMS tokens doesn’t match well to Nahuatl letter order, either, which further weakens the authors’ interpretation of the VMS script as a Nahuatl substitution code.

I’m not entirely opposed to New World interpretations. I think the VMS is probably Old World, but I will listen to New World arguments, as long as they are good ones. Unfortunately, many New World theories are marred by faulty logic and hasty conclusions.

J.K. Petersen

© 2018 J.K. Petersen, All Rights Reserved


Més reflexions sobre 116v…

This might seem a little far afield, but it adds a different twist to reading “anchiton” on folio 116v of the VMS. Think about the fact that the “s” is silent at the ends of words in some of the Romance dialects (which means it wasn’t always explicitly written in medieval text). Note also that “qui” was sometimes written “chi” (as in Greek) or “ki” (as in some of the old French dialects).

Here’s a clip of the famous phrase for reference:

On folio 116v, consider for a moment that the first letter might be pronounced like French “e” (which is nasal), which might be written as “a” by foreigners (substitutions of “a” for “e” were very common in southern Germany and parts of southwest Germany/Alsace). Thus, medieval variations might include

anchiton o la dabas   or   enkiton o la dabas   or   enquiton o la dabas.

Now, following this idea…this phrase is found in the 12th/13th century Crusade Charters :

nous enquitons tous les clains et tous les debas,,…”

We can pare this down to, “nous enquitons les debas,…” or simply, “enquitons les debas”.

The “s” letters at the end of enquitons and les are not pronounced. You could write it enquiton le debas* and it would be understood. In fact, in medieval text, the “s” on “les” was sometimes omitted, with singular and plural written the same. Translated, it means, “ask about the debates”.

*The word “debas” is a medieval spelling of debats (debates), so the “t” was sometimes dropped, as well.

In the region where French and Spanish dialects blended, we could interpret “enquitons o la debas” as “inquiries or the debates”. Or, alternately, since the word “and” was sometimes written “e” instead of “et”, it might become “enquiries and the debates” (this is definitely stretching it since the “o” doesn’t look like “e” and these two letters are not swapped as often as “a” and “e”).

There is also the possibility of “en quiton” and “enqui ton” and, in Tsakonian (a western Greek dialect), εγκι (enki), which is the neuter form of “this”.

So, exploring a nasalized “e” written as “a” opens up quite a few possibilities if “a” turns out to be a dead end.

J.K. Petersen

© 2018 J.K. Petersen, All Rights Reserved


Hiding in Plain Sight

I often wonder if the text on VMS f116v was a later addition to the manuscript, and whether the writer had any knowledge of Voynichese. The Voynichese characters in the lower left are often mentioned but I’ve never seen an assessment of the folio overall. Are there clues that have been overlooked?

Researchers mostly agree that “aror sheey” in the lower left is Voynichese, even if they don’t agree on whether the first glyph is “a” or “o”. There are quite a few ambiguous “a” or “o” glyphs in the VMS, so some transcripts include both possibilities.

However, just because the Voynichese lines up with the other text on the line, I didn’t want to assume there was a connection between the Voynich characters and the 116v text. The ink that follows “aror sheey” is darker, and the letters are Latin, with different thick-and-thin proportions. It isn’t necessarily written by the same person. The drawings and Voynich glyphs might have preceded the other text by days or years.

But is this the only Voynichese on the folio?

In a past blog, I’ve mentioned that the “i” shape in “vix” looks like Voynichese. It’s darker, is distinctively backleaning, and doesn’t resemble any of the other “i” letters on the folio:

pic of the i letters of f116v

I am not sure if the word after oladabas is multos, imiltos, or something else, so I did not include the debatable “i” in this word. I’m also uncertain about letters 2 and 3 in abia/ahia/alka, so I didn’t include this either. All the other “i” letters are fairly consistently upright, with distinctive serifs, as can be seen from the samples. They do not resemble the “i” shape in “vix”:

pic of VMS f116v "vix"Okay, so maybe there is another Voynichese character besides “aror sheey”, but why just one in the middle of a line? When there is only one, it’s hard to speculate on the reason. Was the writer a VMS scribe who slipped into an old habit? Or is it purely coincidence?

Here comes the interesting part, something I haven’t seen mentioned… might there be another Voynichese glyph? One that’s completely overlooked?

Look at these “a” letters. Even though the handwriting is variable, they do have some important commonalities:

The most distinctive traits of these letters are the upright pointed stems, and the distinctively upturned serifs. I’ve tried for years to find handwriting that matches this “a” shape and it’s quite difficult.

Now look at the second “a” in this clip, from one of the most scrutinized words on the folio:

With the angled stem and lack of a serif, this character resembles the Voynichese “a” more closely than the other “a” glyphs on this folio. But is it Voynichese? The shape is right, but the stem stretches slightly farther below the baseline than most Voynich “a” glyphs.

If it is a VMS “a” then it might be important. For example, perhaps

  • whoever wrote the 116v text knew Voynichese and used it occasionally (maybe for pronunciation purposes? or to represent a letter that doesn’t exist in Latin?), or
  • whoever wrote the 116v text was at least familiar with Voynichese glyphs, even if he or she didn’t know the meaning of them, and wasn’t necessarily using the last page as a place to write something completely unrelated to the rest of the manuscript, or
  • whoever wrote the 116v text helped pen the Voynichese text and would sometimes slip into the habit of using the often-repeated glyph forms.

After 10 years of sampling medieval text that resembles the letterforms on f116v, it seems probable that it was written in a late-14th or 15th-century script, so the writer could have been contemporary with the creators of the VMS. If the “a” in oladabas is Voynichese, then it increases the likelihood that the writer was aware of the VMS shapes in a more than casual way and maybe even helped design or pen other parts of the VMS.

J.K. Petersen

© Copyright 2018 JK. Petersen, All Rights Reserved

Tricky Text

Here’s a chunk of cryptic text that looks like Latin at first glance, early Latin, in a 13th-century script. To read it, you need to remember a few things about old-style script…

In the early medieval period, they wrote letters differently…

  • “a” was written like two cees joined—”cc”,
  • “t” was small and round like a “c”,
  • “e” was sometimes written without a crossbar, and sometimes with a longer embellished crossbar,
  • the stem of the “i” sometimes had a slight curve, similar to the rounded “t”,
  • a common style of “r” had a long foot on the bottom (it almost looks like a square cee), and
  • roman numerals with several ones in a row or words with two “i” chars at the end, usually added a descending tail to the last one so it looks like a “j”.

I won’t keep you in suspense any longer. Here’s the cryptic text with a rough transliteration:

The transliteration doesn’t have to be perfect to demonstrate that there’s something a bit weird about this text.

What language is it? It looks vaguely like early medieval Latin, but the common words aren’t there, and it’s somewhat more repetitive than one would expect.

Is it another language expressed with Latin characters?

I’ll let you think about it before I provide further information.

J.K. Petersen

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

Four Play

I have written numerous times about Voynich “4o”, on my blog, on the forum, in comments on other blogs, but I have a feeling it needs to be done again because there may be misunderstandings about how the “4” glyph relates to others in Voynichese and to ciphers in general.

I’ve often seen Voynich researchers say that “o” usually follows “4” and, from a visual point of view that may be true but, in terms of understanding the VMS “4” character, it would be wiser to say that “o-tokens” are frequently prefaced by “4”. The idea that the “4” and the “o” are inextricably linked is not completely true and the idea that they are a unit, in essence one character, is not supported by evidence.

Does 4o Relate to Diplomatic Ciphers?

The diplomatic cryptography systems documented by Tranchedino are based on many-to-one and one-to-many substitution schemes, which means there are sometimes hundreds of characters per cipher system. The cipher on folio 1r of Diplomatsche Geheimschriften packs almost 340 glyphs into a single cipher system.

Think about that… 340 shape-combinations for a single cipher. This is in stark contrast to the VMS, which renders most of the text with 22 glyphs.

Diplomatic ciphers also coded names and common words into symbols and cipher pairs. For example, the cipher glyph for “Papa” is the Latin abbreviation symbol v + “-is”.

This creates a huge demand for shapes, but many were based on a core set. Here are the most familiar characters used to create the cipher system on f1r of Tranchedino’s collection:

It’s clear that Cipher 1r relies heavily on Latin letters, numbers, and abbreviations. Arabic and Hebrew letters (other than kabbalistic loops and lines) are conspicuously absent.

The straight and wiggly lines above some of the letters (not shown) are Latin macrons (the medieval version of the apostrophe). In fact, there are dozens of Latin abbreviations, too many to include on the chart. There is also a smaller percentage of Greek letters and abbreviations, and math/astrology/kabbalah/runic symbols. Sometimes Greek letters are combined with Latin endings (e.g., Greek letter theta + Latin “is” abbreviation), which is not uncommon in Latin texts.

The cryptographers created many new shapes by adding an extra line to a common shape. Or they would remove a stem or add a loop, as in these examples:

pic of letters modified to create new cipher shapesThere are very few shapes that are pure inventions. Most are common shapes, combinations of common shapes, or common shapes altered in regular ways, as in the above examples.

So what inspired “4o” in diplomatic ciphers?

In Latin, it is a common convention to add a small “o” to represent words that end in “o” like “modo”, “quomodo”, “quo”, “libro”, “ergo”, “tertio”, or “quarto”. Latin abbreviations are also applied to homonyms, so go can also represent “gradu”. Thus, we have mo or mo (modo), go or go (ergo/gradu), quo or qo (quo), and 4o or 4o (quarto), etc., several of which are included in Cipher 1r.

One of the more common of these abbreviations, often seen in herbal manuscripts, is “gradu” (grade/degree). It is typically written like the character highlighted below-right.

This is highly abbreviated Latin, so I have transliterated it to make the ordinal number (tertio) and the “o” abbreviation (gradu) easier to see. Both are abbreviated in the same way. Note that the “g” has the “o” attached:

The abbreviation “o” can be placed in almost any position and still mean the same thing. It can be directly above the letter, superscripted to the right of the letter, or attached directly to the letter.

Diplomatic Cipher 1r includes abbreviations for “modo”, “secundo”, and “quarto” (4o). The 4o is not an invented shape. It is used to represent ordinal numbers in several languages.

Related Abbreviations

Here is an example of a related abbreviation (4or ) so you can see that this is a generalizable pattern. Even if you’ve never seen it before, you can figure out what it means because it is constructed in the same manner as mo, go, and 4o:

I chose this example for a second reason… it is written by two different scribes, one using the older form of “4”, the second using the newer. Note the “soft-4” (the older form) and the “sharp-4” (the one that became popular by the mid-15th century). In the VMS we see the “4” character written both these ways, and there is also an in-between shape that resembles “q” more than “4”:

pic - rounded and sharp 4

Pic of VMS 4 that looks like q

This variability could have a number of explanations… different scribes, subtly different glyphs (possibly with different meanings), or scribes from the early 15th century who were transitioning to the newer form (I’ve seen foliation where the scribe changed the older rounded 4 to the newer one and then continued the rest of the numbers with the newer form, as though someone had told him to update the style).

Here is an example of how “quarto” was abbreviated in the 14th century, with the rounder form of “4” that was popular at the time:

The following examples, dating from the 14th and 15th centuries, illustrate how the shape of the four gradually changed from the rounded form to the sharper, more modern form. Note how the sharper form starts to show up in the late 14th century and is more frequent by the mid-15th century:

Picture of medieval numeral 4 and how it changed during the 15th century.

Is 4o one character or two?

In the VMS, “4” and “o” are strongly associated, just as “th” or “ly” are strongly associated in English, but I don’t see anything that points to them being one character. If you write t + h to create a “th” ligature, they are still two different letters. The same applies to 4 + o in languages that use Latin conventions—they are written together but still represent separate characters.

As I have illustrated in previous blogs, the VMS “4” glyph is not always followed by “o”. Sometimes it is followed 1) by other glyphs, 2) by a macron (which, if it were Latin, would indicate missing letters), or 3) by a shape that resembles a small Gothic “l” or the right leg of EVA-k or t:

Another picture I’ve posted before is the long-stemmed glyph that is similar to the “4”. Is it the same glyph or something different?

I’ve updated the sample by adding a date from a 14th-century manuscript that doesn’t completely answer the question but shows how diverse numerals could be in the Middle Ages. It’s not even clear if the marked letter is a 4 or a 9. It’s an oddball numeral that I’ve only seen once, yet it struck me as similar to the oddball character in the VMS, with a shorter stem but a similar swooped curve. At least we know from the context that it’s a number:

Is 4 a prefix that doesn’t require “o”?

There are many “o” words in the VMS and the “o” character sometimes stands alone. It’s possible “4” has an affinity for “o” words rather than “4o” being a character. Perhaps it’s a marker or modifier. Maybe it’s a number.

The 4 char has some interesting properties.  It appears only once on f1v, yet is numerous on f3r. Voynichese is highly repetitive, but it’s not like homogenized milk, there are peaks and valleys, and 4 is not always at the beginning of tokens.

To really understand it, it has to be determined if sharp-4, round-4, and long-stemmed “4” represent the same thing, and it should be remembered that the “o” that follows “4” might relate more strongly to what follows than what precedes it.

J.K. Petersen

© 2018 J.K. Petersen, All Rights Reserve


In his most recent CipherMysteries blog, Nick Pelling zeroed in on a shape on the top line of f116v in the Voynich Manuscript. The letter in question (there are plenty of questions) has been interpreted in more ways than I realized. Pelling has suggested that it might be, “…a rare way of writing a Gothic ‘s’ shape”. I have to admit, “s” never occurred to me when I examined the letter. Not even once.

Here is a snippet that includes the mystery letter (focus on the first character). Underneath, I include a color-enhanced version to make it clear which shape(s) we are talking about and what I see when I look at it.

Pelling says he proposed in 2009 that it might be read as “simon sint (something)”. I found this  puzzling. No matter how I look at it, or split up the pen strokes, I don’t see a medieval “s” at the beginning (I’ll post examples of Gothic “s” further on):

To clarify my thoughts on this…

First, I do not see the first letter in the first word and the first letter in the second word as necessarily being the same. To me, the second one might have a faint descender and a horizontal line just to the right of the descender (under the smudgy part). It’s more squished than the first one (in the horizontal direction). It might be the same letter and it might not. The serifs at the beginnings of words often look similar on different letters.

I couldn’t see any descenders in the multispectral scans, but whether a descender shows depends partly on resolution and partly on which spectra are chosen. The first letter doesn’t appear to have a descender, however. The one on “put?fer” might. The letter on the right word looks vaguely like a “p” but I’m not sure, so most of my comments will be about the first word and the mystery letter on the left.

Sorting out the Letters in the First Word

I usually refer to the first word as “umen” or “umon”, but ONLY for communication, not because I’m committed to any particular interpretation. I have a list of possibilities and I don’t assume it’s a word—it could be a string of characters (e.g., vmçn), or an abbreviation.

The “e/o” letter is indistinct. It could be “o”, “c”, “ç”, or “e” (or something else). When I enlarge it, looks like there might be a couple of pen skips, so it’s possible it is an incomplete “o” (right). Letters 2 and 4 look like “m” and “n”, but I’m not sure about “m” because the humps are different from all the other “m” letters on the folio. Could it be “in”?

Scribal Habits

Before going into detail about the mystery letter, I’d like to point out that whoever wrote this (assuming a specific individual authored most of it) habitually used leading serifs, some of them quite long. It’s possible the writer learned both bookhand (the more formal handwriting) and cursive hand (for rapid writing). There are many hybrid hands with elements of both (see previous blog about the letter “g” which has a bookhand tail and low end-serif).

Here are examples of letters with leading serifs. The serif on the letter “i” is longer than average for scripts of this style:

Now here comes the surprise…

I couldn’t figure out why Pelling kept referring to the first letter as “^”. I assumed he was trying to be neutral about the letter’s identity by choosing a symbol instead of a letter, which is actually a good idea. It was several hours before it hit me that maybe he was interpreting only the serif as a letter. My reaction was, “Whoaaaaaaa!!”

It’s been a week of surprises palaeography-wise. I did not fully appreciate, until the last few days, how differently each researcher sees these characters.

Here are my feelings about it…

The serif at the beginning of the shape on the right is not a letter. If it were, the only typical letter it might be in Gothic script would be an undotted “i” with a very long serif.

An extra-long serif is not  unusual at the beginning of a word, but it still doesn’t look very much like “i”, in my opinion, it looks nothing like “s” either. Also, if the “^” shape were a letter, then what is the blob attached by a stroke on the bottom? The right stroke is not written like the other “i” shapes. NONE of the other “i” letters on the folio has a crooked stem or connects to the previous letter along the bottom. I think this is one letter, not two—one letter with a long serif.

So what letter is it?

You may have noticed that the longest serif of all is on our mystery letter, but is it unusually long? That depends on the identity of the letter. A long leading serif is unusual on the letter “i” but completely normal on “u” and “v” shapes.

Before I post the v/u examples, I’d like to clarify the medieval letter “s” to explain why I don’t think the beginning of the word is “s” (not even a rare one)…

Examples of Medieval “s”

Based on direct observation and sampling thousands of medieval manuscripts, I have identified seven primary forms of “s” in scripts of the same basic style as 116v:

  • straight “s”
  • long “s” (essentially a straight “s” plus a descender)
  • final-“s” sigma (inherited from Greek)
  • final-“s” B shape (similar to modern ß but usually representing one “s”)
  • final-“s” snake shape (like our modern “s”—not common in most countries, although Spanish manuscripts often have this form of “s”)
  • figure-8 “s” (a true figure-8, not one that is deliberately skewed like a cursive “d” or accidentally similar to “8”—this was not common in cursive hands, but is sometimes found in book hands)
  • esszett (commonly expressed on computers as ß, this character had slightly different meanings in different languages but was frequent in central European manuscripts)

The straight-s (which modern eyes can easily mistake for “f”) was more popular in the early medieval period. The stem does not go below the baseline. The long-s has the same hook shape as straight-s plus a descender.

In the early medieval manuscripts, the straight-s was sometimes the only form of “s” used (which means it could be in any position in the word). In other manuscripts, a different “s” (final-s) was used at the ends of words. By the late medieval period, most scribes used a different “s” at the ends of words and some used multiple forms of “s”, as the example below-right:

The straight-s was gradually replaced by long-s. Straight-s is not very common in scripts that are similar to the 116v text, most of them use long-s.

Sometimes scribes added loops or flourishes, but the general form was the same. This chart illustrates that the VMS long-s is quite ordinary:

I’m not aware of any “s” shapes that resemble the first letter or even the first two penstrokes of “umen” but the above forms match well with the first letter of “six”, the last letter of “gas/gaf” and the last letters of “oladabas/oladabad”, “imiltos/multos/miltos” and portad/portas”, so the 116v script is reasonably conventional.


I don’t have a definitive ID for the mystery letter. It looks like the top of an open-p with a long leading serif, but I can’t see a descender (at least not on the first one) or any rubouts under the letter.

It comes closer to a flat-bottomed “v” than the remaining letters of the alphabet but I haven’t found a close match (the flat-bottomed variant is not as common as those with pointed bottoms). The vee on the right is a little too flat—it has lost the “v” shape.

Here is a chart of v/u letters common throughout the medieval period. There are a some flat-bottomed versions circa 1355, 1395, 1400, 1402, and 1410, so it’s possible this style was more prevalent after the mid-14th century, but I haven’t had time to confirm if this is true:

Coming back to the second letter… if this shape is “in” instead of “m”, it might be read as “vinen”, which has meanings in several languages (come, they come, the wine, the vines).

If the last word is “putrifer” then “vinen putrifer” (the grapes ferment/the wine ferments) would be hard to ignore as a possible interpretation. In certain germanic dialects, the “n” at the end of “vinen” is like adding the article “the”.

But what if it’s an “o”? Then it might be vinon or uinon which is harder to pin down than vinen. Vinon is a place in France, but a place name doesn’t seem like the best fit with the other words on the line.

Is anything gained by studying unknown letters?

Even if we can’t make out the letter, the serif on the mystery shape has a calligraphic “brush stroke” feel to it, as does the tail and dipped oval of the letter “g” on the last line. And yet, it’s not professional calligraphy. Maybe these clues hint at other skills…

Was it someone who could draw or who used a brush for some other craft? Medieval artists and illuminators were sometimes illiterate or semi-literate. Perhaps the writer contributed the nose on Aries, painted some of the plants, or inked the secondary breasts on the nymphs. The style of writing is 15th century and might even be earlier in the century if the “a” in 17r “mallier” is a double-story “a”.

I’ve never assumed the writer had any involvement in the creation of the VMS—notes on back pages were often added decades later—but the possibility is there… and that makes it more interesting.

J.K. Petersen

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

Gee, I Never Would Have Guessed!

The VMS marginalia on folio 116v has a number of unclear letters and others that are reasonably clear. Fortunately, a few of them are repeated so we can see variations of the same letter, such as “h”, “i”, “m” and others. For the last decade I have been seeking matches to the marginalia in medieval manuscripts and incunabula, hoping to find the scribe (obviously not a professional scribe, but maybe there’s something out there). I don’t have a match yet, but I have some interesting paleographic data.

It surprised me to discover that one of the letters that I considered clear and readable has been challenged. It has been suggested that the letter following “nim” in “so nim — mich” is “ez” rather than “g”.

I take exception to this. I also do not consider the “plummeting rock” shape after the word “mich” to be the letter “o”, as discussed in my previous blogs.

Here is the phrase in question:

Note that the tall letter with a hook is a medieval “long s”. It’s only an “f” if there’s a crossbar. I read this as “so nim gas/gaf mich” followed by a small drawing.

I can’t tell if the third word is “gas” or “gaf” (both were used in the Middle Ages). There’s an abrasion on the parchment, so it’s hard to tell if it’s “s” or “f” but the letter in question is not the last one, it’s the first one. Another Voynich researcher stated on Nick Pelling’s CipherMysteries blog that the word that looks like “gas” or “gaf” is actually “ez as”. I don’t agree.

Here is a color-enhanced version of how I see it:

It’s a typical “g”, common for the time. The scribe does not write “e” like this and “z” is not typically written like the part on the right side of this letter in medieval scripts, not even as an “ez” ligature. I believe the first letter in the word is one letter and it is “g”. Especially note the serif (the tick on the right).

In medieval scripts that overall resemble the VMS marginalia, the letter “z” usually looks like the shapes in the chart below:

Are there other possibilities?

For the record, the “g” shape is not a medieval “9” abbreviation either. The medieval “9” abbreviation at the beginnings and ends of words was popular for centuries. The “9” abbreviation looks and is positioned pretty much as you see it in the VMS (so I included the VMS “9” glyph along with the other samples in the chart below with the date c.1425 for reference).

Here is how the “9” glyph looks in the VMS. Note that it is positioned the same way as in manuscripts that use Latin scribal conventions, mostly at the ends, but also commonly at the beginnings of words. I’ve written about this many times, but here is a visual refresher:

Sometimes the “9” char was drawn simply, sometimes ornate, but it always signified the same thing in medieval manuscripts… an abbreviation (usually con-/com- or -us/-um).

Here are examples of how the con-/com- abbreviation looks at the beginnings of words (it was essentially the same shape at the ends of words). Note that a serif is expressly not included to help differentiate it from the letter “g”:

So, the marginalia “g” does not resemble a “z” or an “ez” ligature and it does not resemble a “9” abbreviation. It does, however, fit comfortably with common forms of medieval “g”, as in these examples:


There are many shapes in the marginalia that I can’t make out. Some letters have abrasions, some are indistinctly written, some are partly filled in or rubbed out. But I don’t think there’s much ambiguity about the “g”. There’s nothing unusual about the shape or its position in the word.

If someone has a different interpretation for this letter, they can post their paleographic evidence. Personally, I think it’s one of the less controversial letters on the page.

J.K. Petersen

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

Anchiton, Michiton, or An Chiton?

There’s a controversial word on the last page of the Voynich manuscript that is often read as “anchiton” or “michiton”. I’ve written about it before and so have many others, and yet the question hasn’t been settled despite decades of study. I’m hoping some paleographic insights might help.

The troublesome word is on folio 116v (near the beginning of the second line). The individual letters that form “chiton” or “chi ton” are not controversial—they are pretty clear and fairly conventional. Most people agree on them. The only unusual thing I noticed is the extra-long leading serif on the letter “i”. This is a less common way to write “i”:

The extra curve on the “t” is not unusual if the writer learned to write the more traditional round-stemmed “t”. The rounded “t” (written like a “c”) was popular for many centuries, from the early medieval age into the 15th century. Here are some samples of rounded “t” and straight-stemmed “t”  in scripts with some overall similarity to the marginalia writing style:

In the 20 scripts with the greatest overall similarity to the marginalia, both rounded and straight t are represented, but most of them tend to be like the VMS t, in-between the two extremes:

So, putting aside “chiton” for the moment, let’s take a close look at the first letter, or two letters, since it’s not clear whether it’s one or two:

I can understand why anchiton/michiton is contentious. The first couple of letters can be read as ni, an, or mi, depending on the handwriting. Even “mehiton” (vaguely Semitic if it is a hard-h as in mechiton) might be reasonable if the other letters “e” were similar to the character preceding “h”, but they are not. It looks like “ch”.

The problem is further complicated by the less-than-professional-level script—the slants are all over the place, the loops are connected in different ways, and the letterforms are moderately inconsistent.

Nevertheless, I have some observations…

Note that the ending leg of “an” or “m” is not drawn like “i”. This writer has a tendency to draw “i” with a long leading serif, a straight stem, and no ending serif. The last minim on “an” is not drawn this way, so I am inclined to rule out “mi” or “ni” as a reading for this word. That leaves “m” or “an” (or perhaps a very unusual ligature “am”).

What about “an”? Here’s the full passage again, so you can look at all the instances of “a” and “n”:

Note the following characteristics of the handwriting…

  • The letter “a” is mostly tall, with a point at the top of a straight stem, but not always.
  • Ascending loops are usually sharp and at a certain angle, but not always.
  • The figure-8 letter (which is usually interpreted “d” or “s”) usually has a larger bottom loop, but not always.
  • The “n” is usually small and rounded, but not always.

So how do we know whether it’s “an” and both letters diverge slightly (the curve is squished on the “a” and the loops are sharper on “n”), or a loop-m that diverges even more?

Examples of Loop-M

Here are examples of “loop-m”, a particular style of medieval “m” that looks like a ligature-“an” to modern eyes. These are all chosen from unambiguous sources where it can be verified that the shape represents “m”. Loop-m was used more conservatively than regular–m. Some scribes only used it for names or for emphasis:

Did you notice that almost all the samples differ from the VMS in one important detail? Loop-m in medieval manuscripts always has a tail. Always… well, almost always. There are very few exceptions, and even the exceptions tend to have a short tail or a down-pointing end-stroke rather than a serif, in comparison to how the scribe wrote regular “m”.


I stated years ago that I was leaning toward “an” rather than “mi” and it has taken many years to find enough time to explain why. And yet, even though I lean toward “anchiton”, I’m not certain of either reading…

  • If this is “michiton” then the “i” is written differently from all the other “i” characters on the folio, and the “m” is an unconventional loop-m with no hint of a tail.
  • If this is “anchiton” then the “a” is a bit squished, and the “n” is more angular than other “n” characters on the folio.

In fact, I’m not even sure this is one word. It could be “an chiton” or “an chi ton”, which looks suspiciously like an awkward Greek transliteration. It could be coincidence, but if you search the Greek words chυτά/Χυτά or chυτό/Χυτό, and filter for the metallic ones, you will see some very ornate Voynich-like Greek and Russian oil lamps and incense burners:

Images courtesy of Nioras.com and Holy Archangel Liturgical Supply

Medieval versions were probably less ornate than those pictured above (although some of the medieval Jewish spice jars were very ornate), but the tradition of metal censers for funerals, healing rites, and sometimes exorcisms, goes back a long way and the word chytoú for “cast” goes back to biblical times.

If the text on 116v is a healing charm or medicinal remedy (not saying it is, but it’s a reasonable possibility), then a cast/molded burner (chytó, chytón χυτό) for incense (or even as  source of flame for other purposes) would not be out of place.

J.K. Petersen

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

A Stitch in Time

I’ve posted several blogs on hats and tunics and the VMS Gemini tunic is now being discussed in depth on the voynich.ninja forum, so it’s clearly a topic of interest. I’ve been researching the clothing of the zodiac figures for a number of years, so I have many examples from a large variety of sources (including mosaics and stained-glass windows), but I thought I would focus on fashion in two specific manuscripts.

I see the VMS zodiac tunics and robes as belonging together in terms of style.

  • Both male-Gemini and Sagittarius wear basic tunics with simple double-line neckbands, both are wearing hats, Gemini a simple rounded hat, Sagittarius a hat with a very long rounded tail (similar to a foxtail hat, but with fabric rather than fur). Sagittarius has the hint of a goatee. Note that Gemini is conspicuously short-statured even though there’s room to make the leg longer.
  • Both females (assuming the slightly androgenous Virgo is female) are wearing long robes with embellished sleeves and the hint of an undergarment peaking out from under the outer sleeve:

The clothing in Vatican Pal.lat.871 has been mentioned before, because it has many commonalities with the VMS figures.

Here are examples from folio 4r. The subject matter is quite different from zodiacs, it’s a nativity scene, but the roundel-with-text presentation, drawing themes, and clothing have VMS parallels in necklines, gathers, and the hard-to-find bootlaces:

It’s possible that the round-tailed style of hat doubled as a carrying pouch since tunics generally did not have pockets. Small items were strapped to belts, carried on the back, or slipped into hat bands or pouches. The VMS hat does not look like an animal tail. Note that many of the neckbands are similar to the VMS (a plain double-line band), even though a variety of necklines are represented:

There are also tunics with bumpy or scalloped edges like those in the VMS:

One can also find sleeves that are narrow at the wrist and wide at the elbow (left), which is less common than sleeves that are mostly even or much wider at the wrist:

The illustrator was definitely making distinctions in dress. Not all collars were a simple band, there were also high collar, capes, and cowls. Tunics were sometimes single-layer, sometimes double.

There are also many hat styles in addition to the “pouched” hat, including Phrygian hats, royal crowns, tonsured monks, berets, bowlers and, since this is a biblical text, pointed and flat hats to represent Jews and Philistines:

The drawings in Pal.lat.871 make it look like the “pouch” style of hat was common, but it is not easily found in medieval manuscripts. Usually, the tail was an animal tail or the ends were ragged, like a cock’s comb. Pouched hats with very long tails, like VMS Sagittarius, are especially hard to find, although I previously posted this one from Morgan M.453 (left) and one from a Swedish book of law that has a fairly long tail, with a conspicuous roll for the band:

Getting Back to Tunics

Many of the robes in Pal.lat.871 are long or have simple edges, but there are also tunics that are distinctly pleated (e.g., some of the warrior tunics, left) and some that are drawn with a bumpy, gathered, or scalloped edge, like the VMS:

So we can see numerous parallels to VMS clothing styles throughout the manuscript, not just in one or two places.

Finding the Origin of the Manuscript

I was curious about who drew the illustrations.

Pal.lat.871 is written in German, and I noticed it was a dialect. It is thought to be from central Germany, possibly north Hessen (near Frankfurt) or west Thuringia (about midway between Frankfurt and Prague). There is a woodcut version of the Pauper’s Bible created nearby in Bamberg, just north of Nuremberg (c. 1460s) with some of the same clothing styles.

I don’t know if it is specific to the illustrator, but there’s a political statement on folio 16r, a nimbed figure holding the battle banner of the Scandinavian tribes. This puzzling image is sandwiched between Sampson carrying tablets and Jonah in the mouth of the whale. It’s the only roundel on the folio without text.

In manuscript art, the white cross on a red field frequently represents the Lombards or Danes. The inverse of this flag, a red cross on a white field, often represented Helvetians, Templars, or participants in the Crusades. By the time this manuscript was created, Lombardian rule had long since diminished, and Lombardy itself had receded from Florence north to Pistoia, but it still dominated what is now northern Italy, and there were still pockets in Germany, Switzerland, and southern Italy.

But, I have also seen the white-on-red flag in drawings of 14th-century “Gaisler” (Geißler), Christian flagellants associated with the plague years.

Perhaps a sister manuscript can shed some additional light on origins.

A More Primitive Drawing Style

Vatican Pal.lat.1806 was created at approximately the same time as Pal.lat.871 and has very similar clothing themes. It is interesting because the illustrator’s skill level is a little less accomplished than Pal.lat.1806 and thus closer to that of the VMS. Here are some examples of tunics:

There are also sleeves that are narrow at the wrist and wide at the elbow, but they tend to be paired with fancier tunics. Here are some of the simpler ones

Also, if you keep looking, you can find the Sagittarius “pouched” hat. The hat (right) in Pal.lat.871 is not just a vague or generic drawing, it is drawn in a distinctively different way from the ragged-fabric chaperone on the left:

Pal.lat.871 and Pal.lat.1806 are thought to be from different towns, and are drawn by illustrators of different skill levels, yet the clothing themes are clearly related and, in turn, are similar to the VMS zodiac costumes.

Drawing Skill and Cultural Differences

What happens if the same subject matter is interpreted by someone from a different culture and significantly better artistic skills? Do the tunics change? That’s a subject for a separate blog, but I’ll include a few examples to introduce the topic. On the left, from Pal.lat.1806 and on the right, the same scene from BNF Latin 512:

Here some more specific tunic comparisons between Pal.Lat.871 and BNF Latin 512, both of which are from approximately the middle or third quarter of the 15th century:

As might be expected, there are more details in the drawing by the better artist, but paging through the manuscripts side-by-side, there also appear to be small cultural differences that are probably related to the difference in German and French origins. In terms of clothing style and drawing skill, the VMS is obviously more similar to Pal.lat.1806 than BNF Latin 512.


I have much more on this subject, and don’t have enough space to post about the female dress in the same blog. For the moment, there are enough examples to illustrate that the two German manuscripts Pal.lat.871 and Pal.lat.1806 (in addition to those mentioned in previous blogs), bear notable similarities to the costumes of the VMS zodiac characters.

J.K. Petersen

© Copyright 2018 All Rights Reserved