By Gordon Rugg
In this series of articles, we’re imagining that you’ve gone back in time, and that you want to produce the Voynich Manuscript as a hoax to make money. We’re looking at the problems and decisions you’d face, and at the implications of various possible solutions.
The first article looked at why a mysterious manuscript would be a good choice of item to hoax. The second article looked at some of the problems involved in hoaxing a text that looked like an unknown language, from the linguistic viewpoint. The third examined the same subject in more depth, and the fourth discussed the choice of materials, going into some detail about the choice between using freshly-made or already-old vellum.
In this article, I’ll look at issues involved in planning the layout, structure and contents of the book.
Not as exotic as the interior: the outside of the Voynich Manuscript. (Image courtesy of the Beinecke Library)
A traditional book is made as follows.
Suppose you want to make a book of 240 pages. You start by taking 60 sheets of vellum, and folding each sheet in half. Then you gather the folded sheets together into batches of four, with each batch looking something like this. (I’ve coloured each sheet slightly differently for clarity.) Each batch is known as a quire or a gather. Each quire of four sheets will become sixteen pages of the book.
A quire or gather of four sheets
Once you’ve created the quires, you then sew each quire onto a backing strip of strong material, usually cloth, which goes on to become the spine of the book.
One feature of this approach often causes problems for beginners trying to plan a book. This arrangement of sheets means that only one sheet per quire has pages that are numbered consecutively. Here’s an illustration of how that works.
Page numbering in the first quire of a book
The darkest coloured sheet in this illustration gives us Page 1 and Page 2. Then the other sheets give us pages 3 through to 14. We then return to the darkest coloured sheet for Page 15 and Page 16.
This can make it hard for novices to work out which sheets need to contain which text, if you’ve already decided what text the book will contain (as opposed to e.g. writing a journal, where the book is already bound before you start writing in it).
Suppose, for instance, that Page 2 ends with “Take a pinch of” and that you want to run on the sentence to the next page, so that the next page begins with “salt”. You don’t write the word “salt” on the same sheet – if you did, it would appear on Page 15. Instead, you need to write it on the next sheet, which will give you pages 3, 4, 13 and 14.
Why do bookbinders use such a complicated arrangement? One reason is that this reduces the amount of stitching to a quarter of what it would be otherwise, making the book cheaper and also thinner. Another is that it’s easier to write the text and do the artwork on the sheets while the sheets are flat and separate – you don’t have problems from trying to write on a page that’s curving round in a fold where it joins the middle of the book.
So far, this is just traditional book binding. There are variations on this approach, particularly after paper is introduced, but the one described above is the usual one.
So how do you plan what to put in the book and where to put it?
If you’re a novice at this process, one simple approach is to decide which sections you want in the book – for instance, a section on plants, a section on astrological diagrams, etc – and then decide how many quires you will allocate to each section. This means that each individual quire fits neatly within one section, so you don’t need to worry about getting the pagination right for e.g. the transition between the plant section and the astrological section.
The length of each section will then be a multiple of 16 pages (the number of pages in a quire). If your book is going to be 240 pages long, this gives you 15 quires to allocate to the different sections.
This is a manageably small number to handle mentally. You might decide, for instance, that you’ll allocate 8 quires to plant images, 2 to astrological images, 2 to text, and 3 to miscellaneous strange images. If you’re using an established type of book such as an alchemical herbal as your model to imitate, then you can simply base the section themes in your book on the themes in that model.
So far, so good. This works neatly if you’re using freshly made vellum. But what if you’re using an already-old book with some used pages at the start, as a source for already-old vellum to make your hoax look more ancient and authentic? This is where things start to get interesting.
Suppose that you’re using an old book which has writing on pages 1 to 6 inclusive, so the first unused page is page 7, as shown below. The written-on pages are shown as dark reddish-brown.
One possible solution is simply to remove the whole of the first quire. However, that would mean losing almost half a quire of unused vellum.
A more economical solution would be to cut out the used pages, far enough away from the centre stitching to let you re-use every sheet in the quire, and still within the un-written-on margins of those first pages. The diagram below shows what this would look like, with the removed material shown in a darker shade.
If you re-used this mutilated quire without changes, as the first few pages of your book, that would look suspicious; a sceptical viewer would start wondering whether you’d simply cut out the first few pages of an already-used old book. So, if you used this approach, it would be wiser to unpick some of the unused quires of the old book, and distribute the mutilated sheets among them, swapping each mutilated sheet for a complete sheet from another quire. The mutilated sheets would now be spread throughout the book, and it would look as if some vandal had previously gone through and cut out particularly interesting pages, as quite often happens with illustrated books.
If you did this, it could have one irritating side-effect. If the old book had already had its quires labeled for when it was bound, your new arrangement might mess up that original labeling, and cause complications when you came to do the new labels for the quires in your new arrangement.
Interestingly, the Voynich Manuscript shows both these features – some pages have been cut out, and there are signs of the manuscript having been re-bound into a different quire structure at some point in its history.
Is this a sign of hoaxing? The previous interpretation in the Voynich research community has been that the missing pages are simply due to theft, and that the rebinding probably happened because the old binding was falling apart. These are both perfectly plausible explanations, but it would be odd for rebinding to change the quire structure, since rebinders normally try to keep a book together in its original structure when they rebind it. However, a hoax would make more sense as an explanation for these features, since they would be likely side-effects of the method of production.
The planning of the individual pages
One simple way to plan the individual pages is to take some sheets of old parchment (this is for rough sketches, you don’t need anything high quality) and to draw lines on each sheet so it’s divided into subsections. Each side of each sheet gives you eight reasonable-sized subsections. You can then draw a rough sketch of the image for each page and give each page a unique label if required, as shown below.
At about one minute per label for about 120 plants, it would take a couple of hours to give each plant page a unique label at the start, looking as if it was the name of the plant. That’s just the sort of bone that you’d throw to prospective code-breakers; it would look tantalisingly like a real text where each plant page began with a plant name, and codebreakers love names because they’re usually a good way of getting into a code. However, these names are likely to be unlike the rest of the text, partly because proper nouns tend to be linguistically conservative, and partly because herbalists and alchemists often invented their own fanciful secret names for plants, so a codebreaker wouldn’t be surprised if these words turned out to be a dead-end in codebreaking terms.
That’s what we see in the Voynich Manuscript; each plant page begins with a unique word. That’s an easy thing for a hoaxer to do.
However, there’s something that we don’t see in the Voynich Manuscript, which has always been a problem for the code theory and the real language theory. We don’t see anything that looks like cross-referencing, where the first word of one plant page appears in the body text of another plant page. That would be odd, if the manuscript contains meaningful text; you’d expect some pages about plants to refer to other plants mentioned in the same book.
Hoaxing this type of cross-referencing would be easy; you can do it by just adding some cross-referenced terms to the page notes, as a reminder to yourself that you need to add that word among the gibberish on that page. This is shown in the notes for Page 4, in the diagram below.
However, as noted above, we don’t see this in the Voynich Manuscript. That might mean that our hypothetical hoaxer didn’t think of it, but another explanation is that they simply didn’t bother. That’s a theme we’ll encounter more than once in later articles.
A related thing that we don’t see in the Voynich Manuscript involves vocabulary specific to specific sections. If a book has numerous pages which each contain a picture of a different plant, then you’d expect to see repeated mentions of some themes in the accompanying text, such as leaves and flowers and stems. In Culpeper’s Complete Herbal, from 1652, you see all of these terms frequently, as well as mentions of the sign of the zodiac that they’re associated with, and of what each plant can be used for. This is a significant problem for the code theory and the unknown language theory, since it’s a surprising absence; either you need to postulate a code which can hide these terms in some way, or you need to postulate that the textual content doesn’t map directly onto the same themes as the illustrations.
It would be easy to hoax this sort of topic-specific vocabulary, by inventing a set of words and then sprinkling in an arbitrary selection of them onto each page, but this wasn’t done in the Voynich Manuscript. One possible explanation is that we’re seeing a sophisticated hoax, where this was deliberately not done because it would give sceptical examiners too much to work with. Another possibility is that we’re seeing a minimalist hoax that simply didn’t bother.
Text first or image first?
The usual assumption in the Voynich research community is that the images were produced first, and the text was written round them. That’s clearly what happened with the diagrams of the zodiac, where the illustrations contain text within them. With the plant images, however, I found that it was actually slightly easier to produce the text first, and to draw the plants to fit the text.
Here are some examples of different text layouts that I produced, including a couple that were designed to incorporate plants later.
They were easy to produce – it’s surprisingly simple to leave systematic gaps within the text, and the process produces just the same sort of bunching-up near a margin that you get with the image-first process.
There’s one practical advantage in producing the text first. If you’re concerned that something will go wrong in the text production process, then text-first means that if something goes wrong, you only lose the effort involved in the text production, whereas image-first would mean that you would also lose the effort involved in the image production.
It’s long been agreed that the text in the Voynich Manuscript was written by at least two people. The text-first scenario for the plant pages would make sense if one of the people producing the text was something of an unknown quantity, in terms of reliability and likely quality of output. That means less risk of wasted effort than the image-first scenario, if the images were all being produced by the same person.
There’s some evidence within the Voynich Manuscript that’s more consistent with text-first than image-first production of at least some pages. If you look closely at the bottom right of the text in the page below, you see a stalk emerging from the plant stem that suddenly stops just before it hits the text. It looks as if someone was drawing the plant with already-done text on the page, started on the stalk, and then realised that the stalk was going to hit the text. The illustrator would probably be working with the page turned or inverted, so their hand didn’t smudge the wet ink on the previously drawn parts of the plant, so the text might have been temporarily out of sight below their non-drawing hand, if they were using that hand to steady the page.
(Image courtesy of the Beinecke Library)
The same image shows another feature of the manuscript: the technical quality of the images is fairly low. This doesn’t look like a labour of love that lasted for years; it looks like something put together quickly, without much attention to detail. That’s hard to reconcile with the absence of corrections in the manuscript, if the manuscript contains a code, but it’s very easy to reconcile with a hoax where the text contains no corrections because the person producing it knew that it was meaningless gibberish.
The overall planning of a book the size of the Voynich Manuscript can be quite simple, if you’re a hoaxer. Detailed planning of which pages will correspond to which sheets is a more complex problem, if you’re not accustomed to book production.
If you’re using vellum from an already-old book that has old writing in its first few pages, then you’ll need to remove those pages, and this is likely to disrupt the page arrangements, in ways that might explain some features of the Voynich Manuscript.
In the next article, we’ll look at a more low-level issue but more important issue, which is how to produce large quantities of meaningless, complex gibberish text quickly and easily.
The plant clip-art images in this article are from Wikimedia.
There’s more about my Voynich Manuscript work here:
There’s a lot of useful information about already-old vellum and about other fieldwork on Rich SantaColoma’s site:
I’ve also discussed these issues and others in my book Blind Spot:
I’m posting this series of articles as a way of bringing together the various pieces of information about the hoax hypothesis, which are currently scattered across several sites.
Quick reassurance for readers with ethical qualms, about whether this will be a tutorial for fraudsters: I’ll only be talking about ways to tackle authenticity tests that were available before 1912, when the Voynich Manuscript appeared. Modern tests are much more difficult to beat, and I won’t be saying anything about them.
All images above are copyleft Hyde & Rugg, unless otherwise stated. You’re welcome to use the copyleft images for any non-commercial purpose, including lectures, provided that you state that they’re copyleft Hyde & Rugg.