O’Donovan notes #12.4: Rukh, Rocco, Rocca (bird-Merlons).

The author’s rights are asserted.

I’ve decided to post this before, rather than after, taking the month off and to combine what were two 1500-word posts. (Lucky you 😀 ).

Header – added 17th March -all but the lower right detail first introduced to Voynich studies by the present author. That detail, from a small fortress-tower on Euboea introduced by Peter M.

Introduction and Aims

Intrigued by the problem of whether there was some link between the ‘swallowtails’ in western defensive architecture and the form used for the fortress/tower/rook chesspiece, I began from the period of the Piacenza mosaic (11th or early 12thC) and set a limit of about c.1350 AD. Other work had already convinced me that our fifteenth century manuscript’s contents, while they may have contained material first created much earlier, had received their next-to-final version no later.

I began by looking for a linguistic connection between the chesspieces and the ‘swallowtailed’ fortress in other contexts. The key to much pre-modern imagery is not prior drawings, but informing word.

I wanted mainly to test the validity of certain traditionalist habits in Voynich studies.

Knowing that any results from this one, very minor, question and map-detail could not offer definitive answers about when, where or by whom the Voynich map was first created, nor when or where its final recension was made, still I felt that this bit of digging might add clarity to questions about the sense intended for the detail, and perhaps also the range over which it would be reasonable to seek precedents for the Voynich glyphs and what range of scripts and languages might inform a text reaching Latins by c.1350.

For all that, it was a minor matter and I gave it no more than a bare mention in the research-summaries published through Voynichimagery.

Bringing more of the research forward now, I hope it may interest people who like chess and/or who like Beinecke MS 408. I hope too that it may serve to balance, a little, some among the constantly repeated assertions and assumptions made about the contents in the Voynich manuscript.


To begin: linguistic links between bird and fortress.

Here’s one commonly held opinion about chess ..

When the Arabs learned Chess from the Persians, they kept the name rukh, which sounded like the Arabic word for a giant mythical bird… .. When the Italians got the game from the Arabs, the name of rukh was italianized to rocco, which sounded like rocca, the Italian word for fortress.

https //www.chessvariants org/piececlopedia.dir/rook.html

Another way to put this is that when Persian chess-players adopted Arabic, they found no need to change the word ‘rukh’ because it could now be explained using an Arabic etymology. Much the same happened for speakers of Italian, and eventually of English, where rukh became ‘rook’.

The Charlemagne set’s merlons

History often consists of a high, middle and low story, and the same is true for stories how chess came to western Christian (‘Latin’) Europe.

The high story has chess brought directly to the court of Charlemagne by an embassy from Constantinople or from Baghdad. The middle story has it come with unnamed but courtly speakers of Arabic. The low story is that one or more forms of chess were so well known in regions where Christians and Muslims interacted, that knowledge of chess crossed from one to the other by what you might call osmosis.

In this case the high and the middle stories are most likely for early medieval Europe. Most accounts of the dissemination of chess say so, too, and may be summarised as:

During the eleventh and twelfth centuries, after chess had spread to the Levant, North Africa and the Byzantine Empire via the Islamic conquests, chess was played only in noble and royal circles, and so the sets were often made from luxury materials such as ivory and rock crystal. The game is mentioned in writings from the period, notably by Firdausi (934-1020). During the Abbasid period it had been the most popular indoor game played in Baghdad.

The game’s reputation as a game for kings is as old as the legend of its invention by an equally legendary king of India, Shahram.

In the Bibliotheque Nationale de France are what remains of the so-called ‘Charlemagne’ chess set, though it is certainly two centuries later than Charlemagne’s day, and the pieces are generally thought to have been made in Salerno. Despite this, I’m inclined to accept that the embassy from Baghdad whose travails are described by Notker the Stammerer had indeed brought a chess-set if not the one in the BNF, when they came bearing gifts from the Caliph Harun al-Rashid.

The pieces in the BNF are of ivory and so large and heavy (the king weighs 2Kgs) that it has been suggested they served as models in an ivory-carving workshop.

On one we see merlons that are neither square nor swallowtail, but which are attested in Sicily. The same ‘gap-tooth’ design is visible in old photographs of Gibellina, where they adorn a tower but since the photos blur when enlarged, I’ve added another example (a rebuild) from what was a Franciscan friary in nearby Erice. (note that pigeons prefer a lower perch).


Another ivory in the same, Byzantine-influenced, style is attributed again by some to Salerno but by others to Amalfi. This is dated to late in the eleventh century, and shows that merlons of this type are still taken as standard, even while the chesspiece is a ‘swallowtail’ for the makers of that mosaic in Piacenza. We’ll soon see that the ‘swallowtail’ rook/fortress of chess is attested at least as early as the tenth century in Nishapur, and was normal along other parts of the east-west ‘silk’ roads.


The first point to notice. so far, is that the usual practice of calling the detail from the Voynich map a ‘castle’ is perhaps less accurate than to describe it as a fortress or fortified area (It. rocca). True, the chess-rook is today also called, in English, a castle, but the different connotations are important for how the drawing is perceived in the Voynich map.

We have already seen an Italian and a Byzantine ‘winged’ form for the rook, the Italian example a mosaic in Piacenza dated 11th-12thC and the Byzantine piece in ivory to the 12thC. But for newcomers, here they are again.


Having thus passed, with a nod towards Baghdad and Aachen, from Italy, to Byzantium, we’ll now track further along a northern path, widening our temporal and geographic range to do so.

We find, now, that the curious forms given some other pieces in the Piacenza mosaic become easier to understand by reference to a rare set surviving from Persia – from Nishapur – and dated to the 10th-11thC “or earlier”.

FIG 4. sold by Southeby’s from the collection of the late Lothar Schmidt (1928-2013)

Nishapur lies on one of the chief ‘silk roads’ between the Black Sea and China. It is interesting for Voynich research, and for historians of chess, because the city was founded by Shapur 1, the Sasanian-Persian king who kept a Roman emperor a prisoner for life, and to whose court chess is said to have been first introduced from India. Shapur ruled from 240-270 AD.

FIG 5 In the July diagram, the crown in drawn in darker ink, suggesting it a later addition or a later over-drawing to clarify the original.

In representing that moment when chess first came to Persia, a copy of the Shanameh dated 1300-1330 AD shows the rook still has the same ‘swallowtail’ form as in that set and in Piacenza, though the rest of the scene has been re-envisioned to reflect current political reality: Persia under Mongol rule.

FIG 6 https://www.metmuseum.org/toah/hd/shnm/hd_shnm.htm

(Fig. 6 above) “Buzurgmihr Masters the Game of Chess”. Made in Iran or Iraq c.1300-1330. Folio from the First Small Shahnama (Book of Kings) composed by Abu’l Qasim Firdausi (Iranian, born Paj ca. 940/41 d. c.1020 Tus). New York Metropolitan Museum, Accession No 57.51.32

Other physical examples fill the interval between the tenth and twelfth centuries AD. The British Museum holds a twelfth-century set from Nishapur and here, incidentally, one can appreciate why the line of ‘gap-tooth’ merlons might suggest the evenly-spaced ‘little stones’ (It. Gibellina) of a chessboard.

FIG 7 from 12th century. Iran, Nishapur. Stonepaste; molded and glazed. The Metropolitan Museum of Art, New York, Pfeiffer Fund, 1971 (1971.193a–ff) https://www.metmuseum.org/exhibitions/listings/2011/the-game-of-kings-medieval-ivory-chessmen-from-the-isle-of-lewis/exhibition-blog/game-of-kings/blog/carving-out-a-collection

So far as I can discover, no chess-sets survive from Sicily of the 11th- or 12th- centuries, or even from, 13th- or 14th- centuries, but one can hardly doubt that in an area so heavily influenced Byzantine cultural, artistic and religious traditions, together with the centuries of Arab rule, and the noted acceptance of both Muslims and Jews in the courts of Roger and Frederick had seen the game of chess become well-known.

More important for understanding the sense of the detail in the Voynich map is that Rooks of the ‘swallowtail’ sort were known across those roads between Persia and Byzantium. In this summary I won’t repeat the matter I gained initially from fairly obscure sources. Instead, I can refer you to an article I found just a couple of days ago, but which was written in 2013. The next two images I have from that blogpost.


A find from Novgorod – again dated 14thC


It is not difficult to explain why the chesspiece should be associated with that mythical bird the rukh/roc by speakers of Arabic. Ancient sets had an elephant piece where we now have a bishop* and the roc’s chief character in legend is that it is stronger than elephants. In modern chess, a rook/rukh is worth 5 points where a bishop/elephant is worth only 3 – the rook being the stronger just as the rukh/roc was best known as the destroyer of elephants. 😀

*not, as is so often said, in place of the rook.

In western Europe, merlons – even the ‘swallowtail’ type – took various forms, but there can be little doubt that in the west the type consciously associates defensive military architecture with the character of game’s fortress-tower. The piece in the ‘kings game’ bore that significance long before we see the architectural version appear in Italy, associated with the Sicilian-Norman rulers as ‘Gibbeline/Ghibbeline’.

It would be pedantic to begin speaking of them as rukh-merli rather than ‘swallowtail’ or ‘fishtail merlons’ but it would probably be more historically accurate.

For our reading of the Voynich map, the point is that anyone who knew chess-pieces of that form, including non-Europeans, would easily take the form, and merlons in that form, to mean a fortress or fortified structure.

It would be common knowledge among those who travelled or spoke to travellers that some Latins preferred ‘rukh-merlons’ while others, like the Byzantines, used square merlons. And of course if one spoke an Italian dialect, there would be a reflexive association between rocco and rocca. It bears repeating that the fortress-drawing in the Voynich map includes merlons of both kinds.

Within Latin Europe itself, use of the ‘swallowtail’ in drawing could be literal, or might be purely decorative, but one hadn’t to go further west than the Black Sea to know that the Byzantines preferred the one, and the Latins the other, or that in Constantinople (among other places) the defenses in terms of manpower inclued of both Latins and Byzantines.

The history of the rukh-merli is a fascinating sidelight on the way motifs translate between various media and various regions and tongues, but that tiny detail in one roundel in one drawing, cannot tell us where even that detail was first drawn, let alone when or by whom or where our present manuscript was manufactured.

All we can say from the inclusion of those merlons is that the place being indicated was considered heavily fortified. And since my own conclusion after working through the whole map (a task which took an unexpectedly long time) is that the fortress is meant for Constantinople-Pera, it’s fair to mention that the walls of Byzantium were renowned for just that reason.

Postscript to Part 1 (added 17th March). I meant to include the following when speaking of sets associated with Charlemagne; I add illustrations and text as a jpeg to preserve the text.

The roads of chess and of merchants.

Our period of most interest is that century from the early fourteenth to the early fifteenth. The Voynich manuscript was dated to the early fifteenth century by an informed (if anonymous) consensus by Kraus et.al., and this confirmed by a radiocarbon-14 dating of some samples of the vellu which returned an adjusted range of 1405-1438 AD. For reasons explained above, though, our focus is on the century to c.1350 AD.

It may be difficult to accept, but throughout that century, western Christian Europe was no more than a remote marginal area lying at the western limit of the known world. It had almost no importance in the geopolitical scene, even in the Mediterranean where the major powers were the Mongol empire and the Mamluks of Egypt. Constantinople served as middleman in their negotiations and the western trader-states, chiefly Italian, were alternately encouraged or discouraged by the one or the other of the two major powers.

Europe had simply nothing much to offer, apart from its ships and some mercenaries. It is telling that Pegolotti’s guide to the overland route speaks of taking gold and silver coin, but when it comes to trading goods mentions nothing but linens and refers to only one place the trader was likely to find buyers,

Anyone from Genoa or from Venice, wishing to go to [these] places … and to make the journey to Cathay, should carry linens with him, and if he visit Organci he will dispose of these well.

The benefits gained by permitting westerners to pass along those roads was chiefly the benefit of taxation on what they brought, and what they returned with. The Mongol treasury also benefitted by acquiring western silver in the form of coinage, for as that guide says,

Whatever silver the merchants may carry with them as far as Cathay the lord of Cathay will take from them and put into his treasury. And to merchants who thus bring silver they give that paper money of theirs in exchange. This is of yellow paper, stamped with the seal of the lord aforesaid. And this money is called balishi; and with this money you can readily buy silk and all other merchandize that you have a desire to buy.

Nishapur, the old city founded by Shapur I, and where numerous chess-sets from the medieval centuries have been found, was a main hub of those overland routes travelled in both directions by many peoples under the Pax Mongolica and now, for the first time, including some from the Lain west as well as some enthusiastic religious. Though the Genoese and Franciscans appear most often in the historical record, there were some Venetians, and Sienese, and others.

By the beginning of the fourteenth century, the Franciscans’ Rule included a twelfth chapter treating of missions “among the Saracens and other infidels” and assigning their oversight to the order’s Societas Fratrum Peregrinantium propter Christum inter gentes.

We know that before 1350, some Latins had already come to know Amaliq, and that news travelled fairly rapidly from so far to parts of Europe. Frescos made in Siena before 1350 memorialise the execution in 1339, in Amaliq and by the Mongol ruler Jehan Ali, of six resident Franciscans together with a visiting a bishop, an Indian interpreter and a Genoese merchant, the last of whom may have been using their friary as his hostel.

On those Sienese paintings see:

  • S. Maureen Burke, ‘ The “Martyrdom of the Franciscans” by Ambrogio Lorenzetti’, Zeitschrift für Kunstgeschichte, 65. Bd., H. 4 (2002), pp. 460-492. Lorenzetti was active c.1317 to 1348 AD.

I mention Amaliq in particular for a detail seen in a drawing now found on the reverse of the Voynich map and a coin that was circulating in that part of the world from the end of the thirteenth century. This isn’t new information for longer-term readers; it was published at Voynichimagery*, but for those who have come to this blog more recently.

FIG 10

Opinions (as you see) differ on its minting but the point for us is that it circulated before 1350 and is unusual in having the tamgar drawn in a graceful form likely to evoke for a modern, western viewer the French fleur-de-lys.

That coin and comparison to a detail in Beinecke MS 408 was introduced by the present author in January of 2015.

Steve Album attributes that coin’s design to the Taras mint and a date between 1270-1302; Kolbas appaently associates its use with Fars and identifies the tamgar as an unusual form of the imperial tamgar. For reasons explained in my original post, but not repeated here, this coin does not appear to have been used in Fars, even if produced there. However, here is what Kolbas writes:

“”Almaligh produced money in 650 and 651H, and Bukhara and Samarquand issued large flat billon, probably in 651H. …. All of these inscriptions were similar to those of Bulghar and Tiflis in not having the name of the local dynast except in Fars. Instead, they had the great khan’s name and, except in Greater Khurasan and Transoxiana, his tamgha. In Fars, the imperial tamgha was artfully  drawn to resemble a graceful fleur de lys …..from mid-665H” [= 1247 AD].

Judith Kolbas, The Mongols in Iran: Chingiz Khan to Uljaytu (1220–1309) pp.149-50.

Nishapur, Fars, Amaligh and other places named so far in these posts lie along the overland ‘silk’ roads, Fars (mod. Fasa) on the road linking the high overland route to the Indian Ocean and Persian Gulf.

The following map shows the round-journey between the Mediterranean and China, but after 1291, access though the Levant was heavily restricted for Latins. Some of these roads had been travelled by Alexander the Great when he advanced through Persia to the borders of India.

FIG 11

Below – places which have cropped up so far in this blog, in the course of analysing one drawing or another from Beinecke MS 408.

FIG 12 map based on the ‘Silk Roads’ map offered as a pdf by the New York Metropolitan Museum of Art.

On the extreme right of the map above lay Amaliq, where one might have heard Italian spoken early in the fourteenth century.

FIG 13 – courtesy google maps and wiki ‘Amaliq’ article.

Rather than quote from my research sources for the next paragraph, I’ll paraphrase:

Before the mid fourteenth century, Amaligh was a multi-cultural and multi-lingual city. During the so-called Pax Mongolica, Europeans making the journey towards China would stop here. Inscriptions prove that among the various cultural and religious groups found in Amaligh to that time were Nestorian Christians, attested even so late as the third quarter of the fourteenth century.

Giovanni de’ Marignolli speaks of it as ‘Armalec’ and similar forms are found in other Latin works, but as with all place-names, would-be translators have an unenviable task.

When Tughluq Temur became Khan of Moghulistan ( c. 1351) the city became less diverse. Plague soon followed, and when to those factors were added the weakening of the western Mongol-Tatar dynasty, the high overland routes became effectively closed to peoples from the far western edge of the known world.


Who, by the mid-fourteenth century, might be in position to represent defensive walls by merlons of the square and the ‘swallowtail’ type?

Answer: just about anyone with access to the Black Sea or the Mediterraean.

Easterners need have gone no further than the Black Sea (Caffa) to know the Latin ‘Ghibelline’ type resembled the chess rook and no further than Trabizond or Constantinople to know the Byzantine-, square, or ‘Guelf’ type. Both Latins and Byzantines were resident in greater Constantinople and some Latins had interacted with others across the trade routes to as far as China.

A far more telling detail is that sparkling spiral by which the fortress is placed, but that’s another detail, another part of the research into this drawing and so a matter for some other time, perhaps.

To end.. Yet another variety of ‘swallowtail’ merlon.

FIG 14

O’Donovan notes #12.3 the merlon thing. The mapping exercise.

c.2700 words

The author’s rights are asserted.

Update, .

Koen has been in touch, and explained that the project was open collaboration, any forum-members who cared to do so contributing information and pictures they found and these, naturally enough, depended on each member’s decisions about where to look for examples. The map which results is thus a map of those contributions, and indirectly a reflection of members’ areas of interest. At the same time, my impression is that Koen would still welcome a broader data-base. If you care to contribute, but disinclined to have to join an online forum to do it, you might leave a comment under Koen’s blogpost of 2021. [here].

16th March 2023.


Researchers whose attention is focused on the written text in Beinecke MS 408 are looking for single finite answer to what is commonly assumed a ciphertext. I hope I do them no injustice in saying that for most, scientific method is closely associated with statistical analysis, and in recent years we have seen images mapped as data points in an effort to apply forms of frequency analysis.

“Scientific” statistics.

Where in earlier years, the habit among those without specific training in reading pre-modern images was to claim it needed nothing but subjective interpretation and commonsense, or that historical arguments could be made by doing no more than presenting paired images as ‘matches’.

More recently, we have seen some effort made to reduce the treatment of drawings to a kind of statistical analysis which, though not particularly appropriate as methodology seems to be regarded as more efficient than studying the history of art and its methodologies, as if mapping ‘data points’ is by definition more scientific.

Few of us now leave school without learning that any science experiment begins by clearly stating the aim of that experiment. Then by itemising the specimens or samples to be tested. In labs one then describes the method, or at least details of the ISO and Standard Method(s) being applied.

Science students soon learn that if sample-collection is biased; if the methods employed are inappropriate; even if one or more samples are wrongly identified or if the technician is careless in applying standard methods, they still end up with a statistical result but the question posed will not have been answered nor any hypothesis rightly tested.

J.K. Petersen

So far as I’m aware, the first person to try creating distribution-maps for images from Beinecke MS 408 was the Voynichero known only as ‘JK Petersen’. His efforts were hampered by his having evidently little background in medieval history, codicology, or palaeography and none in the history of art or the analytical methods best suited to addressing problematic drawings. His dedication to the ‘all-western-Christian-Germanic-central European’ vision was undisguised and was reflected in his research parameters and thus in his range of processed data.

On the other hand, even within his narrow research-parameters, his work turned up images that certainly assisted others, simply because he introduced so many. I believe he was the first to refer to that French Franciscan manual’s ‘November crocodile’ which assisted our own investigation of the Voynich calendar. What they didn’t and couldn’t do was to prove the manuscript a German product.

Koen and the Lobsters.

Koen Gheuns applied a more nuanced version of that method to clarify one path of dissemination for the ‘lobsters’ in the Voynich calendar, although (perhaps depending on the Warburg database?), he believed the type originated with Michael Scot’s work in twelfth-century Italy and Sicily.

From that point Koen moved forward through northern France to Alsace and to examples seen in images produced from Diebold Lauber’s workshop.

Our subsequent study here complemented Koen’s work by enquiring what precedent works might have influenced Scot’s conception of the Cancer lobsters. We focused on England where Scot received his early education, and France where he received his higher education, as well as southern Spain where he worked for some time and Sicily-southern-Italy where Scot worked in the Sicilian court and where the core copies of his texts were made and first preserved.

Once again, the presence of an effort at data-mapping proved a useful resource, even if it did not prove what the makers believed their statistical-geographic maps proved.

Koen et al. and swallowtails.

Koen Gheuens next approached the topic of ‘swallowtail’ merlons but it is characteristic of Koen’s thoughtful approach that, unlike ‘JKP’ , he began by stating clearly the aim of his latest experiment. His question was “Where were images of swallowtail merlons produced before 1450?”. This was limited by his further aim to more clearly define what is implied when sources speak of ‘swallowtail’ merlons as characteristic of northern Italy. He wanted more clarity on ‘northern Italy’.

The end result was that his research-parameters reduced in practice to “Where, within western Europe, do we find drawings or paintings showing buildings with swallowtail battlements?”

His tacit argument seems to be that wherever we find the most extant instances of such drawings or paintings, that is the most likely place to have seen the origin of this detail in one roundel of the Voynich map (often described as the ‘rosettes page’).

Whether the experiment did – or even could – point to where the Voynich manuscript was made is the question we consider in this post.

As Koen said, in his post of 2021, he had intended to limit the experiment to images in manuscripts and other forms of art. It was a sensible and well-informed decision, but in the event because working with a group he agreed to add a layer marking extant buildings on which ‘swallowtails’ of any type can be seen today.

As a result, the number of data points was greatly increased; the architectural (red) dots largely obscured the iconographic results, and the overall weighting shifted. (see maps further below)

I have tried to contact Koen, first to ask his permission to reproduce his map, and then to ask if he could send a comparison from which the red dots were absent, but so far I’ve been unable to reach him.

Koen assumes that a representation of ‘swallowtail’ merlons will serve as a cultural marker and thus narrow the range in which we might suppose the Voynich manuscript was made. His post of 2021 does not appear to distinguish provenance for the manuscript’s manufacture from provenancing contents.

If it could be shown that, prior to 1440, none but northern Italian draughtsmen created drawings that included swallowtail merlons, or rendered them in wood-carving, relief carvings, mosaics and so on, the chances would be good that the draughtsman who put them on the structure drawn in the map’s north roundel had been a native of northern Italy, or had gained his training there.

But even then, it would be just a fair chance. At the very least one would have to show that all the stylistic details in the map – or in that one roundel – find counterparts in works first created in northern Italy.

And even if the draughtsman had been native to northern Italy, it would not alone tell us where he was when the drawing was first given its form, or when that happened.

Numerous non-Latins travelled in Europe and members of various Italian city-states travelled abroad. Before 1440, we know, some were to be found resident around the Black Sea, in Egypt, central Asia, southern Iraq, Iran and southern China.

I’ll say again – provenancing the manuscript-as-object is work proper to codicologists and palaeographers. Provenancing content is a separate matter – and more exactly, a range of separate matters.

Koen’s range of samples accords with his stated aim of clarifying ‘northern Italian’ in connection with drawings of swallowtail merlons, but the finished map could suggest a certain bias in the sampling. Could – not necessarily does. The reason is poor documentation.

The reader is left uncertain whether the absence of examples from England, France, most of Spain, and the Adriatic (apart from Venice) means that efforts to find examples in those regions returned a null result, or whether the research parameters were so narrowly defined from the outset that those regions were ignored? If the study intended to clarify ‘northern Italian’ it’s understandable, but in that case why include manuscripts from Barcelona and Naples? Was the research heavily dependent, perhaps, on libraries having a large number of their manuscripts digitised?

Koen explained clearly the difficulties involved in adding architectural structures to the data, and in my opinion his initial plan to omit structures was wise, but working as part of a team means compromise. The red dots now swamp the map, and in some cases (such as Genoa), the structures included do not have pre-1440 merlons no matter how energetically civic pride might insist the nineteenth-century reconstructions were historically accurate.

Further difficulties arise because the specimens/data are not labelled, or not labelled accurately in the legend, so that readers are left without any idea of whether the cluster over Milan is the result of a single atelier’s work over, say, 1440-1450, or whether they represent manuscripts made there between 1200-1450. Those placed on Venice may, for all we know, contain a text closely related to that of Naples or Barcelona. Specimen-labelling is basic to any scientific experiment.

Below is a close-up of northern Italy. The whole of Koen’s map can be seen through the link (HERE) which he, and later Peter M., provided.

While this post was in draft, Peter M. directed me to the latest version but it seems little has changed since 2021.

So what does the map tell us?

Not very much. For non-manuscript paintings (the blue dots) none but location details are provided and for the manuscripts (black dots) all the usual information is omitted.

It is impossible, therefore, to determine whether – for example – what we see as blue dots through central Italy is the result of a single painter’s wanderings, or whether due to dissemination of a particular text, or text-type, or the motif’s popularity and transmission within a certain sector of society, such as the intermarrying nobility or a particular religious order.

We cannot follow the chronology for dissemination in manuscripts, since none is identifiable: date, title, holding library and shelf-number are all omitted.

Red Dots.

Koen was perfectly right to urge caution about the overlay of those red dots. Even specialists in military architecture, and archaeologists working in the field are cautious when it comes to assertions about present-day examples of swallowtail merlons.

Even specialists in the history of these forms may have difficulty determining whether some are, or are not authentic reconstructions. I may be mistaken but to the best of my knowledge no original swallowtail merlons are extant in Genoa despite the protests of civic pride that the nineteenth-century rebuilds are authentic reconstructions.

It is true, as in the case of these Genoese merlons (above) that some closely reproduce an early form of such merlons. it is evident from the merlon’s height – able to cover a standing archer – and from the inclusion of slits through which the enemy might be observed and an arrow fired. Nonetheless, the merlons we see today date to the nineteenth-century.

Thanks to a Peter M., who quotes a source he describes as Castle Association and Architecture in the Middle Ages, I can say that before 1450 AD

*The swallowtail pinnacle (merlon) is unknown in the German-speaking part of Switzerland.

*North of the Alps, there are none [no structures with swallowtail merlons?] before 1500.

*From about 1540, castles [north of the Alps??] began to use the dovetailed battlements as a fashionable design. In most cases, even the architect is known.

Context and Significance.

It is understandable that for Voynicheros whose background is in the pragmatic sciences mapping ‘data-points’ might seem a logical way to produce or support any statistical argument, but that expectation is misapplied when it comes to reading and understanding the intention behind a pre-modern picture.

In the same way, people more accustomed to subjects using binary logic have a habit of assuming a division of all images into the literal versus the ‘decorative’, or the meaningful versus the non-meaningful and the same ideas are seen in descriptions of where and why swallowtails were used.

In the critical sciences, data mapping has to involve informed qualitative judgements and more than superficial knowledge of historical periods, documents and cultural attitudes.

Just as counting the number of jars in a medical store-cupboard, and mapping where each stands will not make a doctor of you, so simple data-point maps cannot alone explain he origin, date or intention of any among the Voynich manuscript’s drawings.

The factor most often overlooked is that of perceived significance: in this case what significance the various types of swallowtail merlons had for persons who commissioned, made, copied or regarded a given image in a given social environment and historical moment.

We know that, for a time, one or more types of swallowtail signified ‘the imperial party’. Were the mapped examples found chiefly in cities granted independent status by, say Frederick II, or are they perhaps more often seen in lands that were under the direct control of a western emperor when an image was made? Is the usage dependent a given city’s current political alignment or, perhaps, the leanings of a specific patron? Was a given instance intended to elicit negative or positive response?

Koen’s enthusiasm in 2021 led him to overstatement in his summary, for he wrote:

“The neat thing about this map is that we know for sure that the VM belongs on it as a data point among the blue markers”.

Why should we suppose that for the original maker, those merlons were any more definitive than the square-topped merlons, or more important than the starry spiral, or the various topographic elements? And those are all just in the map’s north roundel. Are any attested in the mapped manuscripts or paintings and if so which?

What lends Koen’s results more weight is that a number of quite independent researchers came to similar conclusions about a focus on the region around, or otherwise connected to Milan, where Koen’s map records the greatest number of manuscripts containing images with swallowtail merlons.

Milan is where Pelling’s historical research, combined with his studies of ciphers, codicology and palaeography, finally led him by 2005-6.

I’ve also concluded, in regard to the Voynich map, that we are most likely to owe its present form to a collaboration of Jews and Genoese, attested not only in Genoa but in Constantinople, Caffa and the Balearics.

Where to from here?

That work done by Koen Gheuens and his friends is not wasted. It should prove very helpful to anyone investigating questions of textual and iconographic transmission in those parts of medieval Latin Europe.

What the map cannot do is tell us where and when Beinecke MS 408 was produced. Establishing a manuscript’s date and place of manufacture is the work of palaeographers and codicologists.

Provenancing content is something else again.

Other questions: Merli, Rook, Rukh.

Etymologies should be taken with a little salt on the side of your plate, to be taken as needed.

For the term ‘merlon’ an etymological dictionary has:

The term merlon comes from the French language, adapted from the Italian merlone, possibly a shortened form of mergola, connected to Latin mergae (pitchfork), or from a diminutive moerulus, from murus or moerus (a wall). An alternative etymology suggests that the medieval Latin merulus (mentioned from the end of the 10th century) functioned as a diminutive of Latin merle, “blackbird”, expressing an image of this bird sitting on a wall.

Let’s start with that tenth century usage, which saw the walls ‘merli’ as blackbirds. Later, in English, the tower-birds, and the chesspiece both became ‘rooks’.

The nineteenth-century etymologists’ well-known disdain for languages other than Latin, Greek and the Germanic group often leaves them blank- when faced with terms gained from Celtic, Hebrew, Egyptian or Berber. In this case the word they cannot see is well known in Hebrew as in Persian – as ruach in Hebrew (esp. Genesis 1:2) and as rukh in Persian.

When working through the Voynich map, section by section and detail by detail, hunting out examples in art and architecture and commentaries in older sources, one chicken-and-egg problem nagged at me.

‘Did the chess-piece inspire the merlons, or did such merlons inspire the form given that chesspiece?’

I scarcely mentioned that part of the research in the summaries published through Voynichimagery, but I’ve decided to write up a little more now for any readers who feel impelled to follow a thread to the heart of an historical maze. That will be in the next post in a couple of weeks’ time.

Until then, here is one detail from a mid-fourteenth manuscript to think about.

The manuscript was made in Persia very shortly before the Plague arose and thus only a very few years before, in France, the ‘November crocodile’ was drawn in grisaille in that Franciscan missal.

As you see, here the rook (Persian rukh) maintains the same form it had in the eleventh century when the chess-players mosaic was made in Piacenza.

Notice in this illustration the hats with stiff, back-turned brims, too, and the garment fastened under the right arm.

Also, from a drawing placed on the back of the Voynich map

a high-collared version..

O’Donovan notes #12.2 – The merlons thing (cont.) Provenancing.

c.1700 words

The author’s rights are asserted.

Part 1 in the current series includes a link to one of Koen Gheuens’ blogposts. If you followed that link you’ll appreciate Koen’s generosity in sharing the work that he and some other members of a Voynich forum were doing.

I meant this post to be about that work and why its approach is exceptional, but because an important point is how their approach differs from most, I think it may be useful to introduce it by considering the problem of provenance-research.

Ever since 1912, Voynich writers have continually confused one kind of provenancing for another, apparently because they did not pause to think through their aims in those terms.

Provenance just means “Where it came from”; the problem arises because of how “it” is assumed defined.

Provenance-research can be divided into three kinds, the first being research into how a finished, or nearly finished object has travelled from where and when it was made to where it is now.

This can be described as ‘chain-of-ownership’ provenance, or – because is primarily associated with descriptions provided by librarians, curators and sellers of artefacts – as ‘Catalogue’ provenance.

That sort of research starts from the time the artefact was manufactured, and ends with the latest acquisition. So in a sense its terminus a quo is finite, but its terminus ad quem indefinite.

Catalogue-style provenance.

The quality of that kind of provenance research can be judged by how severely factual the description is.

To illustrate a near-perfect example of Catalogue Provenance, I’ve chosen that written for a manuscript whose text is written entirely in Tironian shorthand. You will notice that the following description meticulously quotes and dates on palaeographic grounds every post-production inscription (marginalia).

Page from a ninth-century Psalter.

Provenance: [1] A scriptorium in Northeastern France: suggested by the script (according to Bischoff, Katalog (2004), p. 93 (no. 2356)). [2] The Benedictine abbey of Saint-Remi at Rheims: its ownership inscription and book curse added to f. 1r in a 10th- or 11th-century script: ‘Liber Sancti Remigii qui eum abstulerit anatema sit’; its ownership inscription added to f. 1v and f. 33v in a 13th- or 14th-century script: ‘Liber Sancti Remigii Remensis vol VIIxx et V’; the manuscript is listed as no. ‘CXLV’ in an early 13th-century manuscript catalogue from its library (see Dolbeau, ‘Un catalogue fragmentaire’ (1988), p. 215). [3] An unknown 19th-century French owner: added a description of the manuscript in French on f. 101 recto. [4] Thomas Thorpe (b. 1791, d. 1851), London bookseller: [5] purchased from him by the British Museum in August 1832 (see note on f. [iv] verso).

*numerals in square brackets by the present author.

It would obviously be an error to try provenancing the original manuscript by its marginalia, yet in Voynich studies we regularly see efforts made to create a story for the whole manuscript’s content, and for its place and date of manufacture, from no more than a couple of lines of undated marginalia. That’s one of the things which makes Voynich writings seem so very odd to the wider world of manuscript studies. It’s just a wrong way to go about things.

One reason Voynich studies sees such curious habits as attempting to use marginalia to provenance both ownership and subject-matter, is that when interest in the manuscript was revived in the 1990s, the little book written by Mary d’Imperio was adopted as an easy-to-read ‘bible’ by a number of amateurs, many of whom had no prior acquaintance with medieval studies, historical studies, palaeography, codicology or the technical aspects of art history. The same had been true of William Friedman.

Even before Jim Reeds’ ceased managing the first Voynich mailing list, an idea was gaining ground that provenancing the manuscript was a matter of getting an ‘idea’ and then attempting find ways to persuade others to believe that idea plausible. When a television program was made focussing on legends and various theories, the habit of story-telling was reinforced. The people interviewed were not codiocologists or palaeographers, or specialists in medieval history, but persons who had read d’Imperio and who had a novel theory of their own.

You may read a dozens of equally inventive theoretical Voynich narratives today, but it is rare to find any which do not conflate Catalogue provenance with one form or another of Contents provenance.

On the one hand, Catalogue provenance begins with the object’s manufacture and moves forward in time, tracing the hand-to-hand passage of the object from when it was made until now. On the other hand, Contents provenance involves tracking back from the time of manufacture to discover how the text(s), images and materials came to be at last in the place where they came together to make the object/manuscript in question.

So Contents provenance ends at the point where Catalogue provenance begins.

Wilfrid Voynich was the first to conflate the two when he guessed the manuscript made in thirteenth century England, and then relied on nothing but his imagination to assert the whole content of the manuscript created in thirteenth century England, and then interpreted all the contents in terms of what he imagined that thirteenth-century Englishman, Roger Bacon, would write about.

Provenancing Contents

Provenance of this sort rightly asks, “How did these materials and contents come to be employed in the making of this artefact?

Research of this sort, if you think about it, must require require research-parameters and informing sources rather different from those of catalogue-provenance research.

Consider the range, geographic and temporal, needed to get the right answer to such questions as ‘Who composed the Psalter?’ or ‘Who is the King David alleged to have composed some, or all of the Psalter’s contents?’ or ‘What is the origin of the script used in this ninth-century Psalter?’

Provenance research has not been widely understood by Voynich writers since 1912, but chiefly because few stopped to think through their aims in those terms.

Point of View – the drawings

It really doesn’t matter what the modern-day viewer finds easy to understand about pre-modern art. What matters is how the first person to give that image form, and the person who put it in the present manuscript thought, and how they expected their drawing would be understood by their contemporary audience.

For that, it is nonsense to imagine that the modern viewer can pick and choose ad.lib. which images or details they will consider important. If it were true, as it is not, that you can identify the ‘important’ details because (to quote a real Voynich meme) they will be “the most specific and unambiguous” then you’d say the most important detail in this image of King David was his fleur-de-lys crown! But the crown is not unambiguous – you certainly cannot take it to signify that there was ever a King David on the throne of France, or that a French, or indeed an English, king wrote the Psalms.

and so again to the Merlons.

King David. Castile 15th.C

Just as King David lived in c.1000 BC on the other side of the Mediterranean but might be pictured in late medieval western Christian art with a crown as sign of ‘royalty’, and the crown appear variously as one of French-, English-, German-, Persian-, Byzantine- or Spanish type, or with the maker’s idea of a ‘foreign-looking’ crown, so too a structure whose walls had no merlons, or had merlons of some other kind, might still be drawn with those which Voynich writers call ‘swallowtails” and others describe as the Sicilian-Valle d’Aosta type,

Another common error has been to imagine that every drawing is a drawing from life. Yet another has been to imagine that if we find a motif in a manuscript, it is necessarily something copied from some other manuscript.

If we were attempting to research the ‘swallowtail’ motif(s) lineage in art as it might relate to the Voynich map’s examples, we should have to begin from the manuscript’s early fifteenth century date and follow the motif back in time, across a much broader geographic range than is needed to describe the chain-of-ownership and the range of sources and media would have to be broader than Voynich-related writings or only medieval manuscripts.

What Koen Gheuens and his friends did was to carefully frame their question in terms of a specific aim: as he puts it, to discover where, within the Latin west, examples survive of drawings in which merlons are drawn as ‘swallowtails’.

It’s a perfectly reasonable aspect of Content research.

Were the question a broader one, examples would have to be sought in manuscripts and in various other media from the time the first instance of the flat sort of ‘swallowtail’ merlons are attested – round about the eleventh century – until the Voynich manuscript’s date (c.1405-1438).

In that case, the examples would certainly include the next image, though found in mosaic. Dated to the eleventh century, it comes from Piacenza, a town that has cropped up several times in relation to Beinecke MS 408, and first in Reeds’ comments on the Voynich ‘gallows glyphs’ (so called). In this blog, we’ve referred to Piacenza in discussing the assignment of month-to-emblem in the Voynich calendar.

A black ‘swallowtail’ rook is seen in the lower-left hand corner.

Piacenza lies in a region that by now will be fairly familiar to regular readers.

Koen’s post carefully explains too that (a) he had not intended to include extant buildings and (b) swallowtail merlons seen on castles today were not necessarily present, or present in that form, during the fifteenth century.

He was wise to make that point.

The nineteenth century’s ‘Gothic’ revival saw various forms of merlon added to older and to contemporary structures but even examples asserted accurate reconstructions can be problematic.

Take Piacenza’s Palazzi Communale, popularly known as the Palazzo Gotico. The building, or a good part of it, was certainly standing in the thirteenth century, but like most medieval buildings, its architectural history is complex. The Italian wiki [HERE] should provide food for thought.


The subject of Tironian notes has been raised often in Voynich studies. See for example d’Imperio’s The Voynich Manuscript: an Elegant Enigma, and entries to the first (Reeds’) mailing list, Pelling’s book of 2006 or his blog ciphermysteries… for a start.

Reprint – Towers and swallowtails. North emblem and north roundel.

The pictures – all the images appear for me, whether in edit mode or preview or on posting, but it seems that a fair few in this post have not appeared for readers. If you have a problem, just let me know and I’ll reformat them.

The author’s rights are asserted.

What follows is a shorter version of a post from ‘Voynich Annotated News’, October 25, 2021 entitled, ‘Correcting an Identification in my analysis of the voynich map’. It was a summary of work undertaken over several years, and I’m only reprinting this summary-of-a-summary of my research to save time in treating the map’s North roundel, North emblem as examples of the current groundhog-day in regard to the ‘Merlons’ theme.

As ever, you’re welcome to use and quote any of what follows, so long as you allow your own readers to read and evaluate the original from which you’ve taken it. They are entitled to know how information entered the study of this manuscript, and fudging credits denies them their right to weigh and evaluate arguments and evidence. It’s churlish behaviour and the sole cause of this ‘groundhog’ day phenomenon that has developed since the early 2000s. The excuse of theft as a form of ‘checking’ is rubbish. Just direct your chosen ‘second’ to the source you want evaluated. That’s how it’s done.

I’ve cut a little from the 2021 post, but have added a couple more images and notes, re-numbered the Figures and since it’s a very long post, I’ll wait a while before making another post here.


The 2021 post

Apart from hoping it may assist someone work on the written text, I’m reprinting this as a quick way to provide a broader historical perspective on the ‘swallowtail’ motif used in more than one place in the Voynich map.

As I explained when publishing summaries from the full analytical study, walls of such a type had quite a long history in map-making and in that context signified the limit or boundary of empire – all within the empire or the barrier through which one passed into the power of an emperor – and not merely a western or Christian empire.

In the Voynich map, the best known instance has come to be called a ‘Castle’ and is found in the North roundel.

Eventually, after a lengthy period of research and despite a couple of false starts, I was able to conclude from the primary document and numerous other sources, topographic, documentary, archaeological and iconographic, that the maker meant it as a token for Constantinople-Pera and that this North Roundel was part of the map’s final recension, or recensions which I date through the period 1330-1350. I’ll reprise some of that matter later in the post; some of my sources were published with the original study-summaries from 2010-2015 are are not repeated here.

Of especial interest is that during the twelfth century, the Jewish author known as Benjamin of Tudela visited Constantinople and Pera, noting that the Karaite and the Rabbinic Jews lived outside Constantinople proper, in separate, walled, areas and with only a communicating gateway. The area was Pera (Galata) and would then be initially then shared by, and later given into the control of, the Genoese. The following drawings show the area after the Genoese had built (or rather re-formed) its tower.

FIGS 1, 2.

[2023] – I add two more drawings from the full analytical study. The first is an effort at literal landscape yet shows that Constantinople and Pera might be conflated in drawing, even if named separately. The second shows how, in fifteenth century Italy, a token image for Constantinople understands that the ‘swallowtail’ can serve as a generic sign for ‘imperial boundary’.

FIG 3. – French galleys of Captain Polin in front of Pera at Constantinople in August 1544.

FIG. 4 Detail from a copy of Gregorio Dati’s Sfera, National Library of Helskini. Italian, 15thC

I’ll come back to the Pera (Galata) Tower and the North roundel a bit later. At the moment the point is that I think this custom of separate enclaves for Karaite and Rabbinic Jews was one brought by those communities themselves, and while I find no evidence of such customs in Cairo (for example) it gives us added reason to accept the possibility that the same custom had obtained further north in earlier centuries and during what I shall describe very vaguely as the Khazar period.

The ‘North’ emblem

Above that North roundel in the Voynich map we have a detail whose earlier tentative identification is the one I now [2021] want to expand/emend. This emblem used for ‘North’ is quite unlike the other three emblems used for the map’s cardinal directions in being neither impersonal nor pure formality,* but representing a place which was presumably the one furthest north known to the original maker. (Fig. 5)

*the rising and the sinking sun mark east and west; ‘south’ is denoted by a very old sign for the under-world (as the world below the ocean). These three I explained and documented in earlier-published research summaries and won’t repeat it all here.


A casual glance naturally suggests to eyes familiar with the customs of Latin medieval art, the tripartite division of Isidore’s ‘T-O’ diagram of the whole world, but the impression is soon dispelled when the details are considered. I’ve already treated this matter, but it will be useful to repeat the essentials here.

Two arcs of what I’ve called (for want of a better term) palisades denote the boundary of a central area which is divided into three parts with one twice the size of the other two – these being more or less equal in area. The wider world represented by the map is linked to this place by two types of road, or perhaps better, one road and one pathway.

The one is drawn as a wide, high-road, literally a ‘high-way’ because it is shown as being on an embankment, something which suggests to my mind an area prone to flooding or to other natural events likely to make a ground-level road impassable. The smaller and lesser track is drawn, instead, like a rutted cart-track though adorned with pattern found seen elsewhere in the manuscript and evidently serves as a token for for ‘water’ or ‘-of water’. These features are labelled in FIG 6.

FIG. 6

I’ve also drawn a red ring around something which appears to lie at the foot of the embankment, or to have been used (as rubble?) for the embankment’s base. It looks rather like a single ‘swallowtail’ – but whether the maker intended this detail to convey that idea, or whether it’s an accident or drawing or had some other significance, I’m unable to say.

I now have reason, however, to refine my view of the smaller, rutted-looking ‘water’ path, which I had read as reference to some path through marshy ground, or a ford, but more recent information allows me to suggest that it actually denotes a rough path by which women or servants went out to fetch water. We might speak of it, then, as the women’s gate or the servants’ entrance.

The primary document allowed me too to define fairly well the area in which this furthest-known ‘North’ site must have lain, and to suggest that either Old- or New- Serai might be meant. I’ll reprise that matter further below but the point of this post is to add two other possible sites, in the same area, and both of which have been identified as Sarkel. One is now underwater; the other lies relatively close and since debate continues among archaeologists, this other site is diplomatically described as “the Tsimlyansk’s right-bank fortress”.

Thanks to the work of a seventeenth-century scholar named Ivan Satsyperov, we have a description and groundplan of the first site as it was before the the construction of a new dam in the 1950s made of it ‘Sarkel-the-drowned’. In the words of a modern archaeologist it had been “the most perfect of all the known white-stone fortresses of the Khazar Kaganate.” Sarkel is recorded as having been a Khazar capital and bricks found during excavations included some inscribed with what archaeologists call ‘Turkic’ tamgars. It should be emphasised, however, that the language and culture of the Khazars has left little trace in the historical or in the archaeological record; insufficient to allow any coherent argument that (for example) Voynichese is Khazar.

Even by the medieval centuries, Sarkel had lost something of its native character, being provided with walls and battlements in Byzantine style under the advice and direction – as it is thought – of Byzantine engineers.


Here is the seventeenth-century groundplan.

FIG. 8

FIGs 7 and 9 should make clear the site’s position in relation to the medieval Genoese trading ‘colony’ of Caffa, the Genoese presence in the Black Sea becoming rapidly more pronounced and extensive from about 1291, when loss of the old Crusader-held ports in the Holy Land, combined with a deserved antagonism from Mamluk Egypt, had all but lost the Genoese their access to the eastern trade. The high overland road east from the Black Sea offered an alternative route and one which, for a while, they succeeded in having as their monopoly in terms of Latin competition.

Caffa was not the only Genoese holding around the Black Sea but was probably the most important, and there too we still find evidence of ‘swallowtail merlons’ though just who built them is uncertain. Caffa was still, as it had been since the time of the Greeks and Romans, in an area where many peoples, languages and cultures came together.

As most readers will know, it was to be from Caffa, in ships flying in mid-winter for home in January of 1348, that plague would come to both Genoa and Venice.

Note: the wiki map (FIG 7, above) places Trebizond incorrectly. This shown below is correct (FIG 9). The red marker shows the location of the Sarkel(s).

FIG. 9

(For more detailed information about the Byzantine-Persian update to Ptolemy’s Tables, termed the ‘Persian syntaxis’, which emerged from work done out of Trebizond by two Byzantine scholars during the ‘Mongol century’ when Sarai (or Serai) served as the Mongol administrative capital, see source-materials and references in earlier posts to voynichrevisionist).

FIG. 10 Access to the Black Sea required permission from Constantinople and the Latin (western Christian) centres were not much employed until after 1290. For some decades Venetian access was actively opposed by the Genoese, who had gained favoured nation status in Byzantium by their refusal in 1204 to take any part in the Latins’ rape of Constantinople.

I’m not trying to argue that the Voynich map’s ‘North emblem is any careful portrait of any one of these four fortified sites. What I hope to show is that the evidence of the North roundel and of the North emblem are consistent with the historical, pictorial and archaeological evidence and neither contradicts the other [in terms of the map, overall].

In that part of the world, at a time when Latins were trading through the Black Sea after the loss of Acco, and some were taking the ‘highway’ towards the east from this region, the sites of the two Sarkels, and of the Mongol’s New Serai are found established by a river, some certainly divided into three parts, and having a main entrance towards the highway with a lesser path, unpaved than thus rutted, which emerges from an adjacent boundary and leads to the water.

Here is an archaeologists’ reconstruction of what might be termed the ‘Sarkel layout’.

FIG 11

http://sarkel.ru/istoriya/donskie_kreposti_hazarii_byloe_i_nastoyawee_valerij_flyorov/ and see also Mikhail Artamonov

The relation between the two closely-adjacent Khazar sites and the later Mongol site of New Serai can be seen from the following map, though I’ve had to take this on trust from a wiki site, so it should be treated with a little caution.

For information about the two Khazar fortresses, I am indebted to Valerij Flyorov, whose article can be read at ‘sarkel.ru’.

FIG. 12

We know that between Serai and the Genoese-held enclave of Caffa, there was frequent contact even after the advent of Plague and to as late  as 1380-82, four decades later.  

Ciocîltan, for example, speaks of it when arguing that the several treaties with, and substantial concessions made by, the Tatar Il-Khans to the Genoese were largely due to the Tatars having to concentrate on a drawn-out campaign against the Russians.

the third Genoese-Tatar treaty served the same end. Tatar relations with the Genoese in 1381-2 were thus subordinated to Russian concerns and developed very smoothly, as proven by frequent Caffan contact with the authorities in Solkhat and with the imperial capital in Serai.

  • Virgil Ciocîltan, The Mongols and the Black Sea Trade in the Thirteenth and Fourteenth Centuries (Brill. 20120). p.234. 

Serai can be described in modern terms as the capital of an empire, but for medieval Europeans and Byzantines one suspects designation of ‘imperium’ implied antiquity and ancient civilization, such as that associated with Memphis, Alexandria, Rome or Constantinople itself. 

The Voynich map’s North roundel versus its North emblem.

Within the larger part of the Voynich map are various indications which have allowed me to date the map’s final recension and its North Roundel to the first half of the fourteenth century, and thus to the period known as the ‘Pax Mongolica’, the time when first Old Serai, and then New Serai, served as the Mongol administrative capital.

This same time saw the northern, overland routes eastward – the so-called ‘Silk roads’ – which had always been traveled by groups of merchants, finally see Latin faces among the many different peoples who used those roads linking the Black Sea to China.

A guide was written around 1335-1343 on behalf of a Venetian banking company describes the practicalities of travel and trade along one of the northern routes between the Black Sea and China and recommends the trader hire in Sarai a dragoman able to speak Cuman – from which we learn that a language of which today little trace remains was, in the mid-fourteenth century, a lingua franca spoken across thousand of miles.

That reference to Sarai (or Serai) is in Pegolotti’s Pratica della mercatura.

The two SARAIs (SERAIs)

I see today that the wikipedia entry has been greatly expanded and much improved since I first published the post, so now (2023) I’m leaving out much of what I wrote in this section, including several of the maps.

Instead see: ‘Sarai‘ (wikipedia)

When I did the research itself (in the 2010s), I had no idea that there was a film in the making which would require a reconstruction of one of the Serai sites as it had been during the Mongol century. I don’t vouch for the accuracy of the reconstruction, but note her that it also shows a walled area near water, with a lesser track (just visible on the left) down to the river and a high road upon an embankment – just visible to the extreme right. Enlarging the image will allow you to identify it by the white car which is seen on a road that had been a postal Highway even before the days of Alexander the Great. I assume that the various tracks in the foreground have been made for the movie crew’s four-wheel drives. I suspect – though evidence is wanting – that the medieval site is actually the high ground which you see with some sparse vegetation and a broader expanse of white sand or soil visible. Film-crews are not usually permitted to build new structures on genuine archaeological sites.

FIG. 13. Photo credit: macos.livejournal.com

More photos of the reconstructed site, including the ‘beehive’ roofs of a kind we usually associate with hot, dry sites such as Harran and Edessa in Syria, or with Africa.

The two (or three) SARKELs

Koen Gheuens’ current project.

Added note (26th Feb. 2023). What are now usually called ‘swallowtails’ in Voynich-related dicussions are technically described as ‘fishtail’ merlons.

Indirectly I owe to Koen Gheuens my revision of the earlier identification for the North emblem. I was trying to find again the website on which the following structure from Sicily was described as having the oldest known example of ‘swallowtails’ in the Latin (western Christian) world, was said to be known locally as ‘Sarkel’ and had been erected on the foundations of an old Phoenician temple. This was part of research I’d done in 2010-12 and while I never did find the Sicilian site again – it had evidently been since taken down, hunting for the term ‘Sarkel’ led me to the northern sites. Here’s the photo copied from that site. Credits absent, for obvious reasons.

FIG. 14

SARK – ‘ and ;NORTHERN TONGUESjust notes and some musings from a non-linguist’.

Following a storm of three (3) emailed protests about the post’s being so long, I’ve shortened it by more than 1200 words by removing these sections. I haven’t bothered renumbering the Figures. (March 1st., 2023).

BACK TO MERLONS – Asia minor and back to Sicily.

The fortifications shown below are from Semina, off the coast of what was once Lycia. The Turkish government says it was made by Byzantines, though by ‘Byzantines’ they may mean Armenians.

FIG. 15

Other sources describe the fortifications in Simena castle as “erected by the Knights of Rhodes atop earlier fortifications” but no-one knows certainly or attempts to say when that may have happened, or what the precedent fortifications looked like.

For its link to Armenia, I added to the 2021 post a little from an research post which had been published in my first, blogger blog called ‘Findings’.

I include this matter again because the research contributed much to my work on the manuscript and investigation of the Voynich map (between 2010-c.2013).

I found no other reference anywhere in Voynich studies at that time to travellers’ accounts or to merchants and their documents and handbooks. The idea that the manuscript might not be entirely a product of western Christian Europe was treated as laughable and in fact it was not by reasoned argument but by ridicule – because he considered an Asian language possible for Voynichese – that Jorge Stolfi was effectively driven out, and despite Jacques Guy’s best efforts in an article written for the old Times’ educational supplement.

In particular I found no reference made to the Franciscans – and nothing about Odoric of Pordenone, Hugo the Illuminator, Symon Semeonis or John of Montecorvino.. My experiences since then leave me in little doubt that while the content of my research-summaries has gone to swell other Voynichero sites and writings, it will be a rare and marvellous day when original contributions receive accurate documentation by theorists.

However – while Odoric of Pordenone;s account is lived a bit late for us, and Hugo the Illuminator – who set out with Symon Semeonis – died in Egypt on the outward journey, the link with Armenia comes about because we hear from John of Montecorvino, the first Latin Franciscan missionary to China, that part of his preparation involved study in Armenia. We have only one extant source which includes comments by Odoric about his journey and that comes rather late, but it does include a reference to the language of Armenian which, apparently, Odoric had already studied to some extent. (My source here is Yule and the Silk Road Seattle site): .

At Polumbrum [in India], the commander of the ship said to me in the Armenian language, which the rest of the people on board did not understand, that unless we could procure a favourable wind .. he would throw both us and the bones [of Odoric’s deceased fellows] into the sea. … But as the time passed on, and no wind came, I gave one of the bones to our servant, whom I ordered to go to the head of the ship, and cast the bone into the sea; which he had no sooner done, than a favourable gale sprung up, which never again failed us till we had arrived at our destined port in safety.

… I had earlier commented, still in the old blog:

.. Which just goes to show how persistent and conservative are the ways of mariners.

It is a bit surprising to find that (a) Oderic understood Armenian, and (b) that no-one else but the ‘commander’ – pilot – did. And in this case it is unlikely that the pilot himself was Armenian, though Armenian traders figure prominently in some later accounts of the trade routes.

In many cases the Armenians appear to favour the Roman rite even over that of the Greeks, and we hear that one Armenian king, Hethoum II (1266-1307), even abdicated in order to become a Franciscan monk.

from: D.N. O’Donovan, ‘John of Montecorvino and fol. 1r’, Findings, (30th. October 2011).

Montecorvino is said to have served as an adviser to the Sicilian Norman Frederick II before being commissioned to serve as a missionary to China and joining the Franciscan order.

Passing over the oldest, and the non-Latin examples of these fishtail merlons in Sicily, the earliest extant attributed to a ‘Norman’ occurs in what was still a mixed Byzantine-Islamic-Jewish culture in Sicily. (I sent this image to Koen in 2021, but he’d already included Carini).

FIG 15

FIG 17

We know that particular form was no invention of the newly-arrived Norman freebooters, because it occurs as a Persian-style ivory ‘tower’ chesspiece from Byzantium during in the same, twelfth, century. The rook is also called a ‘castle’.

Ornament of that form was best suited to easily-worked materials such as ivory, terracotta and clay, but a need to use bricks and stone seems to me a likely reason that the simpler ‘swallowtail’ began appearing in regions under Latin rule.

Ghibelline‘ merlons

In medieval Latin records, the usual terms are not ‘Guelf and Gibbeline’ but ‘Church/Papal party’ and ‘imperial party’ but we know the slang terms had popular currency in parts of western Europe when strife and partisanship were rife.

I suggest that the Latins – particularly in Italy – had adopted the style and the term from a Sicilian precedent. What I suspect is that the term had been originally pejorative, because the final and most persistent resistance against the Latins had been the mountains near a town called Gibellina.

The mountains themselves are sometimes described as the ‘Monti de Gibellina’ – as for example by Ian Lee in an article that you may not find otherwise helpful but which I happen to have by me. His map labelled Fig.1 – p.3) in.

Ian Lee, ‘Entella: The Silver Coinage of the Campanian Mercenaries and the Site of the First Carthaginian Mint 410-409 bc’, The Numismatic Chronicle, Vol. 160 (2000), pp. 1-66.[JSTOR]

Gibellina’ = Ghibelline.

Though the town of Gibellina (Sicilian: Jibbiddina, Arabic: “little mount” – جبل صغير) was obliterated by an earthquake in 1968, by then it was a “small city and commune in the Province of Trapani, the mountains of central Sicily, Italy”.* This region had offered the strongest resistance to the (Latin) Norman freebooters when they attempted to take the island and for its mixed Arab, Phoenician, and Byzantine character the term ‘Gibbeline’ or ‘Ghibelline’ I’d guess was initially a pejorative in the mainland, the courts of Roger and Frederick II being notorious for maintaining a multi-cultural court and this resulting several times in the rulers’ being formally excommunicated. it is understandable that, on the one hand, the Sicilian style ‘merlons’ would be adopted by Italy’s ‘imperial party’ called ‘Ghibelline’ by their opponents, but equally understandable if the imperial party should proudly adopt the term in daily speech as a matter of pride.

(A variant form of the ‘swallowtails’ occurs in North Africa, and is seen in Almeria).

In the Voynich map, the particular form given the old motif for ‘imperial boundary’* appears to me to maintains that traditional sense and while drawn carelessly, the instances seen in the Voynich map appear to me nearest the Sicilian-and-Latin form. *explained illustrated and documented in earlier posts to Voynichimagery.

Swallowtails and a different ‘White Tower’ – the GALATA TOWER

There is no need to assume such merlons were physically present in Constantinople and Pera, although the prominent presence in Pera of the Genoese and other foreign enclaves makes it possible. In my opinion, as I’ve said, the so-called ‘castle’ in the North roundel is a token for Constantinople and/or Pera. As I’ll explain below, the tower’s form in that case points to a last recension, one including the present North roundel for the first time, to c.1330-1350.

The sites around the Black Sea which served as Genoese ‘colonies’ were, like Pera, ‘imperial’ territory in the same sense that modern embassies, wherever located, are deemed the occupier’s native land. In this case, they were imperial twice over, Pera, Caffa and other such sites being under the protection of both the Byzantine and the Latin emperors.

So the great tower in the North roundel I take to be the Galata – the ‘milk-white” – Galata being an alternative name for Pera, to which both Karaite and Rabbinic Jews had been consigned for centuries before the Genoese were, first, permitted to reside there and subsequently given control of the whole area.

The great Tower of Galata is usually credited, too, to the Genoese, but this is not entirely correct. A Turkish government site provides the following information

FIG.18 The Galata in an early hand-painted postcard,

Galata Tower was first built by the Byzantine Emperor Justinian in 507-508 AD. The Genoese rebuilt the current tower in 1348-49. Today, it is observed that the part of the building up to the third floor has a Genoese character and the other floors have an Ottoman character. 

The tower was later raised (1445-6). In the 1500s, after being damaged by an earthquake it was repaired by an architect named Murad bin Hayreddin. III, with bay windows added to the Turkish addition. Finally, after another fire (1831) two more floors were added and the present form of the roof made.

Thus, the version seen in the detail from the Voynich map agrees best with the form of the Galata tower though the period from 1348-9 to 1445-6.

A white horizontal line [of mortar] can be seen about four or five courses below the lowest line of arched windows, marking the end of the older Genoese levels, and start of the later Turkish work – as seen in the following details.

FIG 19 The Turkish additions.

FIG. 20 The mortar line is most obvious about four of five courses below the bay window on the right

Imperial’ merlons are found still in the old Genoese enclave of Caffa on the Black Sea, but whether used on the walls which the Genoese built for Pera (Galata) we do not know.

Postscript 2023:


O’Donovan notes 12.1: The Merlons thing.

c.2500 words

The author’s rights are asserted.

First, let’s put that detail into physical perspective.

From the Yale facsimile edition, we find that the entire Voynich map measures, overall near enough to 420 mm square. That’s 16.5 inches each side.

Within that map, the roundel containing the detail of interest measures close to 100mm (3.9 inches)

Within that, the structure given crenellations measures just 30mm by 45mm (1.2 x 1.8 inches).

.. for reasons I’ve never seen explained, it’s a habit among Voynich writers to omit or truncate the flanking arcs, thus reducing the detail to the size of an ordinary postage-stamp – about 25mm (1 inch) square.

That’s about the length of your thumb’s top joint. Try taking your finest pen and reproducing it there some day when you’re bored at work. 🙂

What does this tell us about this detail-in-a-detail-in-a drawing?

First, that given the physical constraints on the draughtsman, this detail isn’t likely to be a detailed and literal portrait of any single structure or location, simply because the scale precludes inclusion of enough details to clearly distinguish one structure from all others of similar form within a purely literal genre.

The same constraints tell us that for a given place or location to be clearly identifiable, some of what is included should be there for its resonance in terms of medieval iconographic forms: that is for its existing symbolic or metaphorical hooks.

To put this another way, the constraints imposed by the available space mean that the structure must serve as token for that place and while some elements may indeed be literal, others will not or may not be so.

Consider the difference between what you would produce if asked to present a literal portrait of your local place of worship, as distinct from marking it on a map. In the first place you would need to add far more detail to show the ways in which it differs from all other churches built in the same architectural style, but in the second case, you might just draw one of these:

or you might use a more literal-looking token, detailed only enough that a visitor will connect your drawing to a local building when she or she comes across it – even though the drawing is in no sense a literal ‘portrait’. Like this:

The idea of a ‘token’ image embraces the purely symbolic and the merely generic.

A token can include some literal details (such as the rose window, if your local building has one), but omits all non-essentials and can even include non-literal details – your local building may not have a cross on its door, for example. As you see, this drawing doesn’t show whether the building is made of stone, of brick, of stucco or of wood, which a literal portrait would have to do.

Given the very limited space assigned this detail in the Voynich map, our default assumption (pending other evidence) must be that for the first maker of this detail, each item he did include here had no less weight and importance than any other, regardless of whether one item is drawn in a way generic, symbolic or literal.

It is irrelevant that one or another of the inclusions springs more readily to a modern eye; it is quite as much a mistake to ignore the form of the central tower or the subsidiary towers or the square merlons as the fishtails.

So then, to the degree that the drawing must necessarily be reduced to whatever the maker considered essential elements, what we have is a token, but one hopes a significant token, for the intended place.

For the moment we set aside the various theories which have the Voynich map a description of some poetic, theological or other set of ways. We’ll start from an initial position that the map is not of some otherworld and see how we go.

Given a combination of an intention to communicate information of some kind, and the constrained space available, we now consider each of those features the first maker considered distinctive – even definitive – and suppose further (for the time being) that the place indicated did exist to so late as c.1440.

The structure is placed between two great curving ‘walls’ though it isn’t immediately clear whether those are meant for topographic or for man-made forms. Certainly no defensive walls would be found extending across the lowest point of a very narrow and very steep river valley – a moment’s thought will show you why.

At the front we see a great entry-way opening directly onto what is shown (by a fairly-well known convention in late medieval cartography) as a waterway, shallow and having only one opening. Today we don’t conceive of the Mediterranean as a large shallow ‘bay’, but evidently that’s how the maker viewed it and in fact that’s exactly what the Mediterranean is. By comparison with the open ocean, the Mediterranean is shallow and it does have only one natural opening to the deep sea, through the straits of Gibraltar.

So, without presuming which elements in the Voynich detail convey information by literal depiction and which by symbolic value, consider the remaining items included.

The enclosed area is drawn about twice as wide as it is deep. It is enclosed on three sides by walls.

Inside those walls is drawn nothing but one great tower, apparently round since two others which are square are found outside the back wall and are clearly shown so.

That central tower is evidently distinctive in having three storeys (assuming one window-token equals one storey). The roof is tall and conical, but seems to sit within the tower’s upper edge, which suggests that between the tower proper and the roof is an upper parapet or walkway.

Behind the rear wall, there is placed to our left one of those square towers and this has its top coloured blue, the same pigment used to colour an adjacent area. Where that ends, to the right, is a second square extramural tower, this having its top coloured yellow.

We may suggest, then, that perhaps here the blue is used to denote water, so the front entryway is a water-gate and the tower to the left, outside the walls is a water-tower. If that is what the maker intended, then perhaps the lines of blue which we see following along the top of those curving flanks could indicate aqueducts of the sort so often seen in Roman-era settlements, especially in the near east. But, on the other hand, those lines may only indicate some natural descent of fresh water down steep hillsides. No need to decide yet which the maker intended. The analysis is still in its early stages.

As we look for sites fulfilling our criteria so far, the nature of the less easily discerned will serve as a test for each possible identification.

And now at last we come to the merlons.

There are at least two, and possibly three forms of merlon shown on these walls – the ordinary square merlons and what might be described as two forms of fishtail merlons or, perhaps, an attempt to draw twice the same form of fishtail merlons, but whatever the case, the form given those across the front appear different from those seen on the right side of the rear wall.

As you’ll see (further below) there wasn’t just one form of fishtail merlon, but of course the difference may again be due to the scale at which the draughtsman was obliged to work.

Since we don’t yet know when the drawing was first enunciated, so even if we date our present copy 1405-1438, some effort will be needed to determine by research which elements are employed for their symbolic information and which are more nearly literal within this token. Nor can be even guess, as yet, whether the intended place and structures remained standing beyond 1440 AD, even if they existed to that time.

A common and very typical error in Voynich studies is to begin by assuming that one can identify the place by adopting Mary d’Imperio’s suggestion that its resembled a castle, and then start collecting photographs of such examples of castles having fishtail merlons as exist today, without doing any deeper investigation of the date to which such merlons are dated – in fact many examples seen in Europe today are the result of romantic nineteenth-century ‘renovations’.

Merlons – geographic range.

Merlons of various kinds, including the fishtail type are attested during the medieval centuries from as far north as the Black Sea to as far south as Egypt, and from Asia Minor to western Europe. Some are attested by contemporary writings, some by relics in near-ruins, others by what little still remains in structures often destroyed and re-built since those times.

These facts are well-known to historians and to students of military architecture, and have been reprised and documented often enough in contributions to Voynich studies, that there is really very little excuse one can offer, in 2023, for such misleading assertions as – and I quote:

The swallowtail merlons on the Rosettes** castle and city walls tie the manuscript to southern-German or northern-Italian contexts.

The term ‘context’ avoids saying ‘locations’ while implying it; the substitution provides a loophole, so that in future the theorist can claim the assertion applies to any time when any southern German or northern Italian may have been in any place – including the Black Sea, or Egypt, or somewhere in between such as Constantinople.

But asserting that the type is tied to southern Germany and northern Italy is easily disproven and here again, the work has already been done and more than once since 2010. If readers find no reference to the earlier contributions or to these Sicilian precedents shown below (again) it may be because those researchers have relied too heavily on specifically Voynich-related sites rather than turning to external and more impartial [non-wiki] sources.

The following three images all show buildings in Sicily, and all having their merlons in original style(s), according to our best current information. The first example dates to the tenth century and it is said the merlons which had crumbled over time were accurately repaired; the second example is dated to the twelfth century; the third to the thirteenth century, from which time we see such forms first used by the Franco-Savoyard Challant family* in the Valle d’Aosta, west of Milan.

*They built the famous Fénis castle, among others, and it remained in the possession of the Fénis branch of the lords of Challant until 1716.

Notice the varied forms given these older fishtail merlons in Sicily.

TENTH CENTURY: This part of the tower dates to the period of Arab rule in Sicily (i.e. from 902AD). Before that time, the island had been part of the Byzantine empire. It was gradually re-taken by Christian forces and freebooters in numerous battles between 999 and 1139AD. The Latins who finally took it decided to keep it rather than returning it to the Byzantine emperor, although Byzantine and Arab influence remained strong in the island to the end of the 13thC.
c. TWELFTH Century.
THIRTEENTH CENTURY – Palazzo Corvaja, Taormina.

Nick Pelling’s historical research led him to made a fair case for the present manuscript’s having been made in, or near Milan.

He did begin by expecting all the content would be the original composition of single Latin Christian author who had lived contemporary with the present manuscript’s manufacture. This was in keeping with most theoretical Voynich narratives to that time (2006)

Unlike the creators of many other Voynich narratives, Pelling adopted standard scholarly ethics and used accepted methods, while taking pains to consider the codicological evidence and, as best he could, to date and describe the manuscript’s palaeography. All this in addition to attempting to explain the whole work in the light of his studies of late medieval cryptology.

Pelling was (so far as I know) the first among the Reeds’ list generation of Voynich writers to pay attention to the implications of the script’s “4o” form, while being chiefly interested in its presence in some early fifteenth-century Milanese ciphers.

Pelling read the Voynich map as a city-plan or city-scape rather than a map in any narrower sense.

Unlike many other creators of variants on the traditionalist narrative, Pelling laid out for his readers the course of his own research. He gathered and then presented and cited honestly the full range of precedents and sources he found; he explained his reasoning and the data used to inform that reasoning. He was prepared – within limits – to debate his own findings as few later traditionalists would do, and as some have never done. Even more in keeping with the better type of scholarship, Pelling himself published comments and responses made to his work – the positive and the negative, both. This sort of open-intelligence attitude attracted so many researchers that to just one of the posts listed below he received more than 600 comments.

Pelling represents the last flicker of that energetic, co-operative and actively debating atmosphere which initially gave the first mailing list under Reeds such energy and which led rapidly to numerous new insights still being re-discovered by those living in the present ‘groundhog day’ fog. Thereafter the rise of a ‘believe my theory or else’ and degradation of ethics and standards in the online arenas saw debate and any honest engagement with informed dissent constantly discouraged or disdained by the more ambitious theorists until today one finds little activity of that kind in any Voynich arena.

Whether any of the Voynich research published since 2006 has moved Pelling’s own opinions on any point, I cannot say. It is something which readers must discover for themselves.

Below are linked two of Pelling’s earlier posts. one about the larger drawing and the other about the detail presently of interest, I add a link to a post made by Koen Gheuens in 2017 and, because it tracks the history of this particular ‘groundhog day’, a post made for this blog about 18 months ago. Perhaps after that discussion of Pelling’s contributions to the study I should add that he and I differ on a great many points, especially those invoking one or other of the manuscript’s drawings. 🙂

more on the backstory in an earlier post at this blog:

  • D.N.O’Donovan, ‘Swallowtailsvoynichrevisionist (October 8th., 2021)

In the next post, we’ll move from surviving examples in Sicily to those which begin to appear from about the 13thC in the Franco-Savoyard Valle d’Aosta.

The Valle d’Aosta is not on the Venetian side of Italy, but in an area which will have become familiar to any readers who laboured through our analyses of the calendar’s ‘July’ and ‘November’ emblems. That is, a region between Milan-Genoa and adjacent to regions in which forms of Occitan were spoken during the middle ages – including Genoa. In the map detail shown below, the marker for Valle d’Aosta is seen slightly left of centre at the top of the image.

Valle d’Aosta in the mountains west of Milan, above the Lombardy plain.

O’Donovan Notes #11 .. despite howls of derision.

2300 words

The author’s rights are asserted.

I’ve commented recently on the fact that we do not know that the Voynich manuscript was made in western Christian (i.e. ‘Latin’) Europe, and since the vast majority of theoretical narratives, and assumptions made in analysing the written text, have taken a Latin origin as one of their unquestioned ‘givens’ the question deserves attention – attention which it would normally have received as the first question addressed by any potential buyer.

Among the arguments I’ve seen offered in that cavalier/commonsense tone which is so common in Voynich writings are that the manuscript is on membrane; that the ink is an iron-gall ink; that the writing proceeds from left to right and other comments which show that the speaker has never looked into the question of whether or not these things are characteristics uniquely European.

None is.

There is clear distinction between manuscripts inscribed with the quill, as against those inscribed using a pen. There are regions in which quires are more commonly composed of four, as against five or more bifolios, and regions where the binding-style is characteristic of a given period and region. The use of sewing-supports is characteristic of Latin and of Armenian manuscripts, but the current binding of the Voynich manuscript is a little problematic, as we’ve discussed in an earlier post. And apart from the use of flax rather than hide for its sewing supports, we have to make clear a distinction rarely recognised by Voynich writers – that is, between when matter in a manuscript is first enunciated, when a given volume containing that information was manufactured, and the amount of time which elapsed between quires’ inscription and their being bound together.

In some cases these periods may be short, while in others the gap between enunciation and the current inscription of a text, and between the inscription and binding the quires into a single block, may be hundreds of years – and the addition of an external cover can occur later still.

None of this information is new to codicologists or paleographers, specialists in medieval history or iconographic analysts, but one sees little thought spared for such things when a Voynich theorist begins hunting support for his/her ideas. Wilfrid Voynich again set the model. He thought the manuscript made in the thirteenth century, so he looked for a single thirteenth-century ‘author’.

Manuscript cultures from the pre-modern era – for convenience described here as prior to the mid-sixteenth century – have more in common than is realised by most Voynich writers.

Let another scholar speak to the point. Advocating development of a comparative approach in manuscript studies, here’s what Beit-Arie writes:

One may marvel at the force of the regularity and continuity revealed in the basic structures, production techniques, social, artisanal and intellectual functions, and the aesthetic principles embodied in mid- and late medieval codices throughout book civilisations in all cultures.

Be they codices inscribed in Latin, Greek, Arabic, Persian or Hebrew scripts, or in the less widespread Syriac, Coptic, Glagolitic and Cyrillic scripts – they all partake of the very same anatomy of the codex: common writing materials, similar proportions and formats, the analogous molecular structure of quiring achieved by the folding and stitching together of a regular number of bifolia, and the use of various means of markings in the margins ensuring the correct order of quires and bifolia.

The great majority of these codices would be set for copying by the laying out of the writing surface and by its ruling in a variety of techniques, most of them shared, functioning as a scaffold for the writing.

  • Malachi Beit-Arié, “The Advantages of Comparative Codicology: Further Examples,” in Jörg B. Quenzer, ed., Exploring Written Artefacts: Objects, Methods, and Concepts, vol. 1 (Berlin: de Gruyter, 2021), 395-404. (now available at academia.edu)

And here we come to another of the Voynich manuscript’s unusual characteristics. The bifolios do not appear to have been ruled out.

In some cases, where a text-block (a set of quires stitched together) has later been trimmed down during the course of binding or re-binding, one may find that if the ruling out was done by pricking, the prick-marks are lost, but this does not appear to have occurred in the case of Beinecke MS 408.

In some cases, the ruling-out was later erased, as it commonly was in Italian humanist works attempting to emulate the minimalism of Greek manuscripts – but some trace normally remains and none has been noted, so far as I’ve seen, by any informed commentator.

Other methods, which used a frame and wire leave some impression on the membrane, but neither have those signs been noted by anyone who has observed the manuscript at first-hand.

Such marks may exist, but from what has so far been said of the manuscript, the Voynich quires show, in their lack of ruling out, divergence from that norm found across all those manuscript cultures listed by Beit-Arié.

This doesn’t prove the manuscript was or wasn’t made in Latin Europe, or in some other region where Latin presence and influence was to be found but it offers a useful research question if you’re looking for one: When, where and in what context do we find membrane inscribed without ruling-out? taking c.1440 AD as end-date.

Form and Content

The habit of ignoring the manuscript’s form, while expecting to define the manuscript’s content by guesswork, statistics or by creating imaginative-hypothetical storylines is as old as Voynich studies itself. Among the innumerable bizarre meme-laws which circulated before Lisa Fagin Davis became publicly involved with the online Voynich community was one which advised newcomers that it was “unnecessary” to pay attention to Nick Pelling’s early emphasis on codicology … because it was “too complicated”. No, I’m not kidding. The quote marks really mark quotes among the many pronounced as if by authority and so so staggered me as a newomer in 2009 that I began compiling a list of them from that time.

In much the same way, it was asserted until relatively recently that the drawings’ interpretations “were all subjective” and, more to the point, that they could only be illustrations for the written text, because that was imagined normal for western Christian manuscripts and it was supposed commonsense to believe no other origin possible for either the first enunciation of the content or for the present manuscript’s manufacture.

On the importance of balancing evidence from the manuscript as material object, and from its contents, I’ll again let another scholar speak. To what follows, a couple of very minor adjustments have been made, but the original paper is linked below so you can check that.

Since a manuscript is a material artefact that is usually produced in order to preserve and transmit specific contents, its role within a manuscript culture must be considered in a twofold way: In terms of its content and in terms of its physical – that is, its material and visual – characteristics.

By content we refer to the information that is encoded in written texts, but also in images as well as other possible sign systems within a manuscript such as musical notation.

The physical characteristics comprise anything from the manuscript’s format and measurements, to the materials chosen as support for writing and painting, to the visual organisation of a manuscript, its decoration and the style of the script.

Content and its concrete physical instantiations can be, and have often been, conceptually distinguished and considered in isolation from each other, but in the production and use of a manuscript they are inextricably interrelated. The intended content of a manuscript is an important factor in determining what physical characteristics the manuscript producer(s) will choose.

(..and vice versa!! A manuscript’s physical characteristics indicate where, when and by whom the content was intelligible. – D.)

Perhaps now readers who have not troubled much with the issue of how we read evidence embodied in a manuscript’s physical characteristics will realise why such things matter and provide limits for the exercise of what is politely called historical imagination in Voynich studies. In passing, let me also commend the codicological studies by Wladimir Dulov.

It is not only a written text which is encoded, and just as imagination alone will not hand us the key to the way a written text is encoded or enciphered, so neither will it provide the key to the information embedded or encoded in the manuscript’s physical evidence – its materials and its drawings.

In fact, just as the material object of a manuscript may contain folios gathered across a range of time, and the content of the written text can reflect sources first enunciated many miles and many centuries apart from one another, so too the drawings have to be studied for signs of diverse origin and what I’ll call chronological strata.

As you’ll see by roaming through what is said and written online about this manuscript, there is a general habit of maintaining Wilfrid Voynich’s assumption that everything in the manuscript had a single author and was composed at the same time the quires were inscribed. For this idea there is no justification except, perhaps, the obvious fact that the manuscript’s appearance is so unlike the expected western-Christian-medieval that commentators defaulted to imagining an amateur its imagined Latin author.

In more recent times we have seen the work recognised as a compendium (one positive advance in the study overall during the past thirteen years). Very recently we have seen Fagin Davis confirm that the written text is not by one scribal hand, but by several. The same had been observed earlier by Currier, and that observation refined by Pelling as early as 2006, and further defined and refined now by Davis in her presentation to the zoom conference in 2022.

In theory, of course, a number of scribes might all have been copying sections from a single master-work, but against this is the fact that the drawings do display what I’ve called chronological layers, and distinctly different – diverse – iconological codes inform drawings between one section and another, to which we must add what appear to be late additions in darker ink by a hand which I should think a Latin Christian’s.

Some Voynich writers such as Reeds, Pelling and Neal have appreciated the importance of codicology while failing to test the long-established theoretical history for the manuscript. Others have simply treated everything but their own variation of the traditional story as negotiable and in this, I’m sorry to say, adherents of the German-central-European idea have been worst. When was pointed out by Pelling that one or more of the scribal hands appeared to be influenced by the humanist style – and humanist style does not appear in the north until later than the vellum’s date-range, so the manuscript’s dating was arbitrarily altered to suit a time when humanist script was introduced to German-speaking regions.

If that theory wanted to claim the leaf-and-root section spoke to Frederick II’s interest in alchemy, or that the containers in that section were German Christian ritual vessels, proponents would simply ignore the physical and historical limits set by the physical evidence or – more often – ‘adjust’ them by asserting that the date could be extended from the early fifteenth century to as much as the mid-sixteenth century.. or that if the vellum were too coarse for German manufacture that German makers could produce coarse vellum too, or that the vellum wasn’t really coarse at all – the latter idea accidentally lent support by the new Beinecke scans’ being so bleached that the evidence of follicles and roughened surfaces was virtually erased and with it – by the way – evidence of the palette’s yellow wash. The manuscript’s curious green stars became blue and so on. To those with only digital access, which is all but a very few, discussion of various folios where that pigment appears thereafter seemed nonsensical.

Very recently, after two palaeographers declined to support a theory of German ‘hands’, the grapevine says that the ‘central European’ theory is about to be tweaked and re-formulated yet again, now to become a German-Venetian-military-Dominican-Franciscan theory that will incorporate (with or without credit) work done by non-supporters of the Germanist theory, and I fear bring to the mix even a work of such appalling bigotry that the religious order whose medieval member wrote it passes it over in silence and in this is followed by most medieval historians, some of whom may mention its title but then pass over its contents for shame.

Still, these rumours about the Germanist theory Mark ?? are still only rumour and worth no more unless it turns up in public.

NEXT POST – more about the manuscript’s drawings and how theorists’ determined erosion of ethical standards and methods from the early 2000s turned Voynich studies into what Pelling once called the ‘Voynich groundhog day’ – when exactly the same work is being done over and over, with each newcomer left in the dark about all previous work and how many began from exactly the same assumptions, followed exactly the same flawed approach, and ended with the same result – an idea that the work supported a preferred theoretical narrative.

Our example for this phenomenon will be a single, very small detail found within one detail within the Voynich map. Detached from its context, this detail has been covered, re-covered, re-discovered and re-explored, by methods re-invented and re-applied and with the same basic errors for almost twenty years.

Until then, perhaps you’d like to apply your critical thinking to the following. It’s a series of isolated comments strung together as if it formed an historical argument reached by an impartial survey of the manuscript’s material and iconographic evidence.

Precisely these same assertions have been offered and repeated since, at least, 2008 and by so many that I see no reason to name-and-shame the latest person to have failed to carefully think about what they chose to believe. Here’s what that person wrote:

The most important illustrations are those that are most specific and unambiguous, and which can offer clues to provenance and/or subject matter.

Many aspects of the manuscript offer clues to geographical provenance in the southern German or northern Italian cultural regions. There are four instances of marginalia in an unknown German dialect. ‘rot’ and ‘r’ appear in plant roots (4r, 29r), and it seems unlikely that a non-author would make such annotations.

The final word of the charm on 116v is ‘maria’ with a superscripted cross between ‘a’ and ‘r’, which—in addition to the cross on 79v— indicates a Christian context.

The Zodiac illustrations bear well-known parallels with southern German manuscripts, which do not need repetition here. The crown on 72v1 resembles crowns of the Holy Roman Emperors and other Austrian royals, including a c. 1350 reliquary bust of Charlemagne possibly made for Charles IV; an archducal crown on a painting of Rudolf IV, duke of Austria (d. 1365); the imperial crown buried with Friedrich III (d. 1493)* observed using an endoscope; a coin of 1484 depicting Sigismund, archduke of Austria; and a woodcut by Albrecht Dürer depicting Maximilian I (d. 1519).

The swallowtail merlons on the Rosettes** castle and city walls tie the manuscript to southern-German or northern-Italian contexts, where such merlons predominated. Swallowtail merlons also appear in documents made in early fourteenth-century Venice, 1340s Zürich, Sankt Peter an der Schwarzwald in 1487, Nürnberg in 1493, and another catalogued as 1300s ‘probably German’.

*The person described by that writer as ‘Friedrich III ‘ (d.1493) is better known to historians as Frederick III, and is to be distinguished from the Emperor Frederick III of Prussia, and both from Frederick III, Elector of Brandenburg.

** [ note expanded 22 Feb 2023]. ‘nine rosettes page’ is a term coined by Nick Pelling in discussing a drawing subsequently analysed in detail by the present author over the period from 2011-2014 and which was shown to be a map. Overall, the conclusions of that analysis -were incompatible with any all-Latin-Christian or any German-imperial-sixteenth century theory (as the latter then was), and to this day supporters of the latter refer either to Pelling’s term ‘rosettes page’ or credit one of the several subsequent authors who attempted to create some alternative map-related interpretation which would conform to a Eurocentric theory. As was said in my earlier post on this motif, Mary d’Imperio, in her book The Voynich Manuscript: An Elegant Enigma (1978) p.21 used the term rosettes and says the structure “resembles a castle”.

Such habits are why the Voynich manuscript’s study devolved so rapidly, from about 2010, into a series of groundhog days unless concerned with statistical and linguistic analyses of the written text. Little attention was paid to the manuscript’s codicology after Pelling’s work, until the Yale Facsimile edition (edited by Clemens) was made available in print.