Voynich proto-optics…?

by ⋅ 17 Comments

I’ve been reading up on the pre-history of the telescope recently (hence my reviews of Eileen Reeves’ Galileo’s Glassworks and Albert van Helden’s The Invention of the Telescope), but have omitted to mention why I thought this might be of relevance to the Voynich Manuscript.

The answer relates to Richard SantaColoma’s article in Renaissance Magazine #53 (March 2007) with the title “The Voynich Manuscript: Drebbel’s Lost Notebook?”, which claimed to find a persuasive familial similarity between the curious jars arranged vertically in the pharma sections and Renaissance microscopes, specifically those described or designed by Cornelius Drebbel. His (updated) research also appears here.

The biggest problem with Voynich hypotheses is that, given 200+ pages of interesting stuff, it is comparatively easy to dig up historical evidence that appears to show some kind of correlation. In the case SantaColoma’s webpage, this category covers the stars, the hands, braids, caps, colours, four elements, Francis Bacon’s New Atlantis and handwriting matches suggested: none of these is causative, and the level of correlation is really quite low. All of which is still perfectly OK, as these parallels are only presented as suggestive evidence, not as any kind of direct proof.

It is also tempting to use a given hypothesis to try to support itself: in the 1920s, William Romaine Newbold famously did this with his own circular hypothesis, where he said that the only way that the manuscript’s microscopic cipher could have been written was with the aid of a microscope, ergo Roger Bacon must have invented the microscope. All false, of course. Into this second category falls the “cheese mold”, “diatoms” and “cilia” of SantaColoma’s webpage: if these are to used as definitive proof of the presence of microscopy in the VMs, the level of correlation would need to be substantially higher. But these parallels are, once again, only presented as suggestive evidence, not proof.

Strip all these away, and you’re still left with the real meat of SantaColoma’s case: a set of striking similarities between 17th century microscopes and the curious devices in the Voynich Manuscript’s pharma section. Even if (as I do) you doubt that all the colouring on the pages was original (and upon which some of SantaColoma’s argument seems to rest), it’s an interesting observation.

Having said that, no actual proof or means of proof (or disproof) is offered: it is just a set of observations, resting upon a relatively little-tested tranche of history, that of the microscope. Can we do better? I think we can.

Firstly, modern telescope historians (I’m thinking of Albert van Helden here, though he is far from alone in this respect) now seem somewhat dubious of the various Janssen family claims: and so I’m far from comfortable with placing the likely birth of the microscope with the Janssens in 1590. As Richard SantaColoma points out, Cornelius Drebbel is definitely one of the earliest documented microscope makers (from perhaps a little earlier than 1620, but probably not much before 1612, I would guess).

Secondly, it is likely that the power of the lenses available for spectacles pre-1600 was not great: Albert van Helden calculated that a telescope made to della Porta’s (admittedly cryptic) specification could only have given a magnification of around 2x, which would be no more than a telescopic toy. I would somewhat surprised if microscopes constructed from the same basic components had significantly higher magnification.

Thirdly, the claimed presence of knurled edges in the VMs’ images would only make sense if used in conjunction with a fine screwthread, to enable the vertical position of an element along the optical axis to be varied: but I’m not sure when these were invented or adapted for microscopes.

All in all, I would assert that if what is being depicted in the VMs’ pharma section is indeed microscopes from the same family as were built by Drebbel from (say) 1610 onwards, there would seem no obvious grounds for dating this to significantly earlier than 1610: even if it all came directly from Della Porta, around 1589 would seem to be the earliest tenable date.

The problem is that there is plenty of art historical data which places the VMs circa 1450-1500: and a century-long leap would seem to be hard to support without more definitive evidence.

As always, there are plenty of Plan B hypotheses, each of which has its own unresolved issues:-
(a) they are microscropes/telescopes, but from an unknown 16th century inventor/tradition
(b) they are microscropes/telescopes, but from an unknown 15th century inventor/tradition
(c) they’re not microscopes/telescopes, they just happen to look a bit like them
(d) they’re not microscopes/telescopes, but were later emended/coloured to look like them
(e) it’s all a Dee/Kelley hoax (John Dee was Thomas Digges’ guardian from the age of 13)

Despite everything I’ve read about the early history of the telescope and microscope, I really don’t think that we currently can resolve this whole issue (and certainly not with the degree of certainty that Richard SantaColoma suggests). The jury remains out.

But I can offer some observations based on what is in the Voynich Manuscript itself, and this might cast some light on the matter for those who are interested.

(1) The two pharma quires seem to be out of order: if you treat the ornate jars as part of a visual sequence, it seems probable that Q19 (Quire 19) originally came immediately before Q17 in the original binding.

(2) The same distinctive square “filler” motif appears in the astronomical section (f67r1, f67r2, f67v1), the zodiac section (Pisces, light Aries), the nine rosette page (central rosette), and in a band across the fifth ornate jar in Q19. This points not only to their sharing the same scribe, but also to a single (possibly even improvised) construction/design process: that is, the whole pharma section is not simply a tacked-on addition, it is an integral part of the manuscript.

(3) Some paint on the pharma jars appear original: but most seems to be a later addition. For example: on f99v, I could quite accept that the palette of (now-faded) paints used to colour in the plants and roots was original (and I would predict that a spectroscopic or Raman analysis would indicate that this was probably comprised solely of plant-based organic paints), which would be consistent with the faded original paint on the roots of f2v. However, I would think that the bolder (and, frankly, a little uglier) paints used on the same page were not original.

Put all these tiny fragments together, and I think this throws doubt on one key part of SantaColoma’s visual argument. He claims that the parallel hatching inside the ornate jar at the top of f88r (the very first jar in Q17) is a direct indication that it is a lens we are looking at, fixed within a vertical optical structure. However, if you place Q19 before Q17 (as I believe the original order to have been), then a different story emerges. The ten jars immediately before f88v (ie at the end of Q19) all have vertical parallel hatching inside their tops, none of which looks at all like the subtle lens-like shading to which SantaColoma is referring. For reference, I’ve reproduced the tops of the last four jars below, with the final two heavily image-enhanced to remove the heavy (I think later) overpainting that has obscured much of the finer detail.

This is the “mouth” of the top jar on f102r: the vertical parallel hatching seems to depict the back wall of a jar, ending in a pool of faintly-coloured yellow liquid (probably the original paint).

 

This is the mouth of the bottom jar on f102r, which appears to have vertical parallel hatching right down, as though the jar is empty near the top (or perhaps its contents are clear).


This is the top jar on 102v, enhanced to remove the paint. I think some vertical hatching is still visible there: it would take a closer examination to determine what was originally drawn there.

This is the bottom jar on f102v, again heavily enhanced to remove paint. Vertical hatching of some sort is visible.

The Book of Soyga, revisited…

by ⋅ 6 Comments

It’s a nice historical detective story, one kicked off by John Dee, Frances Yates‘ favourite Elizabethan ‘magus’ (though I personally suspect Dee’s ‘magic’ was probably less ‘magickal’ than it might appear), when he claimed to have told an angel that his “great and long desyre hath byn to be hable to read those tables of Soyga“. Dee lost his precious copy of the “Book of Soyga” (but then managed to find it again): when subsequently Elias Ashmole owned it, he noted that its incipit (starting words) was “Aldaraia sive Soyga vocor…“.

However, since Ashmole’s day it was thought to have joined the serried, densely-stacked ranks of long-disappeared books and manuscripts, in the “blue-tinted gloom” of some mythical, subterranean library not unlike the “Cemetery of Lost Books” in Carlos Ruiz Zafon’s novel “The Shadow of the Wind” (2004)…

Fast-forward 400 years to 1994, and what do you know? Just like rush hour buses, two copies of the “Book of Soyga” turn up at once, both found by Deborah Harkness. Rather than searching for “Soyga“, she searched for its “Aldaraia…” incipit: which is, of course, what you were supposed to do (in the bad old days before the Internet).

It is a strange, transitional document, neither properly medieval (the text has few references to authority) nor properly Renaissance. There are some mysterious books referenced, such as the Liber Sipal and the Liber Munob: readers of my book “The Curse of the Voynich” may recognize these as simple back-to-front anagrams (Sipal = Lapis [stone], Munob = Bonum [Good], Retap Retson = Pater Noster [our Father]). In fact, Soyga itself is Agyos [saint] backwards.

But what was the secret hidden behind the 36 mysterious “tables of Soyga” that had vexed John Dee so? 36×36 square grids filled with oddly patterned letters, they look like some kind of unknown cryptographic structure. Might they hold a big secret, or might they (like many of Trithemius’ concealed texts) just be nonsense, a succession of quick brown foxes endlessly jumping over lazy dogs?

  • oyoyoyoyoyoyoyoyoyoyoyoyoyoyoyoyoyoy
    rkfaqtyoyoyoyoyoyoyoyoyoyoyoyoyoyoyo
    rxxqnkoyoyoyoyoyoyoyoyoyoyoyoyoyoyoy
    azzsxbqtyoyoyoyoyoyoyoyoyoyoyoyoyoyo
    sheimasddtguoyoyoyoyoyoyoyoyoyoyoyoy
    eyuaoiismspkfaqtyoyoyoyoyoyoyoyoyoyo
    enlxflfudzrxxqnkoyoyoyoyoyoyoyoyoyoy
    sxcahqczfbtfzsxbqtyoyoyoyoyoyoyoyoyo
    azepxhheurgmyknqnkoyoyoyoyoyoyoyoyoy
    rlbriyzycuyddpotxbqtyoyoyoyoyoyoyoyo
    ryrezabirhdiszeknqnkoyoyoyoyoyoyoyoy
    ogzgfceztqalpntsxhssyoyoyoyoyoyoyoyo
    opnxxsnodxqhuekknykkoyoyoyoyoyoyoyoy
    rcqsfueesfsqrqgqrossyoyoyoyoyoyoyoyo
    roauxmdkkxkhyhmpzqphdtgtguoyoyoyoyoy
    aqxmudiamubkoqifbszktdmspkfaqtyoyoyo
    sazoesrmlrnaqnzhgabmsmlpeahfsddtguoy
    ………………………………
    (etc)

Jim Reeds, one of the great historical code-breakers of modern times, stepped forward unto the breach to see what he could make of these strange tables: he transcribed them, ran a few tests, and (thank heavens) worked out the three-stage algorithm with which they were generated.

Stage 1: fill in the 36-high left-hand column (which I’ve highlighted in blue above) with a six-letter codeword (such as ‘orrase‘ for table #5, ‘Leo’) followed by its reverse anagram (‘esarro‘), and then repeat them both two more times

Stage 2: fill each of the 35 remaining elements in the top line in turn with ((W + f(W)) modulo 23), where W = the element to the West, ie the preceding element. The basic letter numbering is straightforward (a = 1, b = 2, c = 3, … u = 20, x = 21, y = 22, and z = 23), but the funny f(W) function is a bit arbitrary and strange:-

  • x f(x) x f(x) x f(x) x f(x)
    a…2, g…6, n..14, t…8
    b…2, h…5, o…8, u..15
    c…3, i..14, p..13, x..15
    d…5, k..15, q..20, y..15
    e..14, l..20, r..11, z…2
    f…2, m..22, s…8

Stage 3: fill each row in turn with ((N + f(W)) modulo 23), where N = the element to the North, ie the element above the current element.

For example, if you try Stage 2 out on ‘o’, (W + f(W)) modulo 23 = (14 + 8) modulo 23 = 22 = ‘y’, while (22 + 15) modulo 23 = 14 = ‘o’, which is why you get all the “yoyo”s in the table above.

And there (bar the inevitable miscalculations of something so darn fiddly, as well as all the inevitable scribal copying mistakes) you have it: the information in the Soyga tables is no more than the repeated left-hand column keyword, plus a rather wonky algorithm.

You can read Jim Reeds paper here: a full version (with diagrams) appeared in the pricy (but interesting) book John Dee: Interdisciplinary essays in English Renaissance Thought (2006). The End.

Except… where exactly did that funny f(x) table come from? Was that just, errrm, magicked out of the air? Jim Reeds never comments, never remarks, never speculates: effectively, he just says ‘here it is, this is how it is‘. But perhaps this f(x) sequence is in itself some kind of monoalphabetic or offseting cipher to hide the originator’s name: Jim is bound to have thought of this, so let’s look at it ourselves:-

  • 1.2.3.4..5.6.7.8..9.10.11.12.13.14.15.16.17.18.19.20.21.22.23
    2.2.3.5.14.2.6.5.14.15.20.22.14..8.13.20.11..8..8.15.15.15..2

If we discount the “2 2” at the start and the “8 8 15 15 15 2” at the end as probable padding, we can see that “14” appears three times, and “5 14” twice. Hmm: might “14” be a vowel?

  • 2 3 5 14 2 6 5 14 15 20 22 14 8 13 20 11 8
  • a b d n a e d n o t x n g m t k g
  • b c e o b f e o p u y o h n u l h
  • c d f p c g f p q x z p i o x m i
  • d e g q d h g q r y a q k p y n k
  • e f h r e i h r s z b r l q z o l
  • f g i s f k i s t a c s m r a p m
  • g h k t g l k t u b d t n s b q n
  • h i l u h m l u x c e u o t c r o
  • i k m x i n m x y d f x p u d s p
  • k l n y k o n y z e g y q x e t q
  • l m o z l p o z a f h z r y f u r
  • m n p a m q p a b g i a s z g x s
  • n o q b n r q b c h k b t a h y t
  • o p r c o s r c d i l c u b i z u
  • p q s d p t s d e k m d x c k a x
  • q r t e q u t e f l n e y d l b y
  • r s u f r x u f g m o f z e m c z
  • s t x g s y x g h n p g a f n d a
  • t u y h t z y h i o q h b g o e b
  • u x z i u a z i k p r i c h p f c
  • x y a k x b a k l q s k d i q g d
  • y z b l y c b l m r t l e k r h e
  • z a c m z d c m n s u m f l s i f

Nope, sorry: the only word-like entities here are “tondean”, “catsik”, and “zikprich”, none of which look particularly promising. This looks like a dead end… unless you happen to know better? 😉

A final note. Jim remarks that one of the manuscripts has apparently been proofread, with “f[letter]” marks (ie fa, fb, fc, etc); and surmises that the “f” stands for “falso” (meaning false), with the second letter the suggested correction. What is interesting (and may not have been noted before) is that in the Voynich Manuscript, there’s a piece of marginalia that follows this same pattern. On f2v, just above the second paragraph (which starts “kchor…”) there’s a “fa” note in a darker ink. Was this a proof-reading mark by the original author (it’s in a different ink, so this is perhaps unlikely): or possibly a comment by a later code-breaker that the word / paragraph somehow seems “falso” or inconsistent? “kchor” appears quite a few times (20 or so), so both attempted explanations seem a bit odd. Something to think about, anyway…

"Das Voynich Blog" (reading German can be easy sometimes)

by ⋅ Leave a Comment

Here’s a link to Elias Schwerdtfeger’s very interesting “Das Voynich Blog”.

Elias has worked really hard behind the scenes to find ways of visualising the statistics expressing what “old hand” Voynichologists (such as, say, Philip Neal & I) see when we look at the Voynich – you know, the highly bonded, multi-level internal structure that exists at the stroke, character, glyph, digraph, word, line, paragraph, page and section levels.

As an aside, I’ve long disagreed with Renaissance encipherment hypotheses for the VMs based on moving alphabets, specifically because they fundamentally destroy these kinds of internal structure: the only way to keep such hypotheses alive is then to argue (as, for example, my old friend GC does) that these structures are part of the “surface language”, i.e. that the encipherer is dynamically stretching his plaintext to mimic these structures in the ciphertext. Yes, it’s possible, but… put all the pieces together and it’s a bit too much of a stretch for me.

Incidentally, I’ve been looking at f2v recently, specifically because of the “fa” marginalia there (one of the very few marginalia I didn’t really cover in my book, “The Curse of the Voynich”). Elias discusses f2v at some length, proposing the eminently sensible (and testable) hypothesis that the same pe(rso)n that/who made the dubious (o)ish(i) emendation to the last line of f2v also added the “fa” marking above the second paragraph. They’re both in similar darker ink (which is a good start): but I think that the Beinecke’s scans – though fantastic for most purposes – fall just short of being able to resolve this kind of question definitively.

Actually, I’ve got a list of about 50 similar/related cross-indexing questions like that I’d like to address (say, by multispectral imaging or Raman imaging) in the future. But for now, that project is stalled (because the Beinecke turned my proposals down). Oh well: maybe next year…

More Voynich-related stuff…

by ⋅ 2 Comments

Here are some current links to amuse (and possibly inform) you…

(1) A 49-second clip on Voynich manuscript page f2v (yes, really!) from that well-known research institution, YouTube. And no, I don’t know what the commentary is saying either (please tell me if you do):-
http://www.youtube.com/watch?v=20o1DSiiiuE

(2) Here’s a wonderful page from Uncyclopedia, an open-source Wikipedia parody (I say “an”, casually assuming that there must be more Wikipedia parodies out there). Yet the text is surprisingly fresh and on the money, and (some would say) no less enlightening or inspiring than any of Gordon Rugg’s media interviews (probably):-
http://uncyclopedia.org/wiki/Voynich_Manuscript

(3) In November 2006, a Swede with the online handle “Pemel” announced that he might have translated the Voynich Manuscript, and that he had passed his solution on to a referee for external checking. OK, so his chance of success is sub-10%, but who knows?
http://www.moopy.org.uk/forums/showthread.php?t=38615

(4) Finally, here’s a link to the blog of “Leandra Evals” (an anagram, surely?) who thinks there’s something to do with alchemy in the VMs. Could be… (though I suspect probably not):-
http://voynichalchem.blogspot.com/