I’ve just finished reading @MargalitFox’s excellent book “The Riddle of the Labyrinth“, which untangles the skein of history around the decryption of Linear B to reveal the quiet (but huge) contribution made by Alice Kober.
Fox’s belief (which I largely agree with) is that Kober would, had she not died early, almost certainly have completed her decryption programme before Michael Ventris. Regardless, Ventris had spent years making a fool of himself by insisting loudly and at great length that the language of Linear B must surely be Etruscan (it was actually an early form of Greek, Δ’Ω!), and he only began making swift progress once he took Kober’s results on board.
Because Linear B was an unknown language written in an unknown script, Kober always insisted that anyone who took a theory about the language as their starting point was doomed to failure. Rather, the single route to the finishing line was, she asserted, to find the patterns and deep symmetries inside the primary texts that we have, and to work outwards from there.
Kober’s attempts to systematically comb through the Linear B texts were frustrated through the 1930s and part of the 1940s by Sir Arthur Evans’ refusal to release more than a modest fraction of them. However, she built up card indexes and added physical cross-referencing means (using carefully punched holes, she was able to optically find matching patterns, like using postcards to build her own Google search facility for Linear B ).
It is easy to draw a long list of comparisons between her sustained attack on Linear B and The World’s somewhat scattershot attacks on Voynichese. Perhaps unsurprisingly, few of these cast the latter in anything like a favourable light.
Similarities
Both Sir Arthur Evans and the Voyniches had very fixed (and, in retrospect, quite wrong-headed) ideas about the historical sources of their respective scripts / languages: and both released only a small number of images to scholars before their respective deaths.
Hence the constraints Alice Kober was working within during the 1930s and 1940s weren’t really so different from those that Voynich researchers ‘enjoyed’ for most of the 20th century. Her specific response was to make her own transcriptions, build her own analytical machinery, and construct her own decryption methodology.
If you want a direct apples-to-apples comparison, I’d perhaps suggest looking for the methodological parallels between Kober and Captain Prescott Hunt Currier (1912-1995). They both consciously and deliberately attacked their targets without a specific plaintext language in mind; discovered deep language-like patterns that nobody had either noticed or grasped the significance of; and then disseminated them openly.
Differences
The #1 difference is that while Linear B had Michael Ventris, Voynichese has had no Gary Lineker or Filippo Inzaghi hanging around on the goal-line to head Captain Currier’s critical cross into the goal.
While it’s easy to say that Ventris was brilliant, in many ways his whole approach to Linear B had been naive and self-defeating from the start. Margalit Fox concludes that Kober thought Ventris was yet another hacky Linear B amateur, far more of a research liability than a research asset: that he was so blinded by his idiotic Etruscan theory that his research would never (in fact, could never) produce anything of genuine value.
But Ventris’ key personal asset turned out to be that he had, as the famous US entrepreneur/investor Fred Wilson put it back in 2016, strong views weakly held. That is, once Ventris finally twigged that Kober had found something genuinely telling that was incompatible with his (previously strongly held) Etruscan theory, he had the strength of character to be able to jump ship completely. (Though admittedly Ventris did strongly hedge his initial description of what he had come up with by describing it as something that might be no more than a wonderful delusion.)
For me, the oddest thing about Voynichese is that even though modern researchers now know a vast amount about its inner workings (for example, you could hardly fault Torsten Timm’s diligence and persistence), they remain steadfastly unable to figure out the next step forwards.
If you can imagine a Voynichese football hanging in the air in front of goal while all the strikers are squabbling at the opposite end of the playing field, you’re not far off the truth. 🙁
Synthesis
Even though many now know about Linear B, what is less known is the story of Linear A. Also discovered by Sir Arthur Evans, the Linear A script is almost certainly a syllabary that was used on Crete to write a (now-lost) Minoan language. When the early Greek invaders came from the Mycenean mainland, they adapted Linear A as a script to write down (admittedly somewhat imperfectly) their Early Greek language.
Alice Kober realized early on that despite the many visual similarities between their sign shapes, Linear A and Linear B were writing down entirely different languages. Hence she abandoned all attempts to decrypt Linear A (because there were so few examples of it) and focused instead on the much more promising Linear B.
In many ways, we have a closely analogous situation with Voynichese, in that it comprises the two major ‘languages’ that Captain Currier identified in the 1970s. More recent research has identified even more subtlety to Currier’s A vs B division: the researcher Glen Claston (Tim Rayhel) asserted that he had identified the specific sequence by which Currier A evolved (or was actively mutated) into Currier B.
Even now, however, it remains absolutely the norm for researchers – even otherwise very good researchers – to carry out their analyses on the whole of their Voynichese transcription, i.e. all the A pages and B pages merged together into a single whole, as if they were all the same kind of thing.
Perhaps it should be no surprise that, to me, this is akin to mixing Linear A and Linear B into a single Linear corpus, superficial amateurish nonsense that Kober had nothing but disdain for in the 1930s and 1940s.
Hence if you genuinely want to be the Michael Ventris of Voynichese, I would suggest that you start by trying to learn from Alice Kober and Captain Currier:
- Assume you know nothing at all about the unknown language(s) beneath the unknown script (because you don’t, you simply don’t)
- Tackle one corpus at a time (say, Herbal A, Quire 13, or Quire 20)
- Build up what you consider to be a reliable transcription for it
- Build up contact tables
- Begin with the patterns at the start, middle, and end of words
- Determine the precise internal logic of the script, with the idea of working out how that might be interfering with the unknown language beneath that script
*like*
Alice in wonderland?
One essential and simple point that you make is that one needs to be prepared to change one’s mind about anything and everything in response to evidence justifying such a change. This is something, as you point out, Ventris did to his credit.
I don’t think there is a problem having formed an opinion such as Linear B is Etruscan. One will, and it is natural to, form hypotheses which may or not be overturned, in fact I think it is necessary to form working hypotheses, as long as one is prepared to change one’s mind.
I think, although I don’t know Ventris’s work well, he was lucky to have investigated a pattern related to city names, which just so happened to be crucial to decipherment.
Nevertheless I would caution against assuming that the experience of deciphering Voynichese will necessarily be anything like that of deciphering a natural language like Linear B. One will only know what the key to deciphering Voynichese was with hindsight and it may or may not be what is expected.
I read recently that it has been argued that as a result of finding more tablets of Linear A, as is expected, it to may well be deciphered in the future; though I don’t know what the hope is for the Phaidos disk and other examples of Cretan hieroglyphics. There are a few other examples of undeciphered languages such as Elamite from Iran.
Mark: Kober (and several others) had considered place names as a possible way in for a decade or so, so that wasn’t in itself a new idea. All that Ventris did was look at the endings in a slightly different way from Kober, and that was enough (when combined with place names) to get a small number of letters. And then all the rest of the dominoes that Kober had carefully set in place toppled. 🙂
Mark, there are two chief problems about the theory of forming ‘working hypotheses’ about material which is incomprehensible to the researcher concerned.
The first is purely practical – if you experience what feels like a sudden blinding insight it is instinctive to suppose you’ve had a moment of brilliance and then to turn, though you may not be aware of it, to a hunt for material to justify that impression. The more rational procedure – the norm in any developing science – is first to observe carefully as much as you can, to then set it out as rationally as you can, testing and double-blinding (or double sampling) to test your observations and once you think they are at the factual level, to then – if you must – begin to generalise *about the facts* to form an initial hypothesis and then eventually one hopes to form an all encompassing explanation – a theory.
The other problem is psychological. When a theoretical model for research involves the creation of discardable working hypothesis it sounds reasonable but it is a trick of human psychology that our sense of self becomes intertwined with that hypothesis, and people will (as they have often done in Voynich studies) soon reach a level of emotional attachment which has them behave as if their ‘theory’ – guess/hypothesis – were in the same relation to them as a wife or child: love me, love my ‘theory’. Intellectual objections to the theory are treated as if they were attacks upon the vulnerable and beloved; the issue – and thus response – become wholly personal to the point where even to ask details of the reasoning, or any textual reference are treated as if to answer them was in itself a loss of face. Some artists feel like that about their creations, as writers often do about their works. Scholars are trained – or supposed to be trained – to have more invested in learning, and to express thanks when an idea is shown historically impossible or a statement inaccurate.
I agree with Alice – first observations and data, and this then checked, stress-tested and double-stress-tested as if you were your hardest and best-informed critic.. then see if you can explain it all with a working hypothesis. 🙂
Nick: That’s interesting. The point I was trying to make is that it may be one specific thing that by chance leads to the decipherment and predicting what that specific thing is in advance may be very hard. So it is likely that there will be some luck involved in stumbling on that one feature or set of specific details that unlock everything else. So Kober may or may not have spotted that pattern if she lived longer we can’t say. Certainly there is something to be said for being in the right place at the right time in research terms. The solution could come via some kind of crib or some statistical observation or some linguistic observation or some completely different means and knowing in advance which line of research will be most productive could be near impossible, though we can certainly have our own ideas. It could be that the decipherer is not the smartest person, but just much luckier than most in just happening to find the key to everything else, like a metal detectorist finding a hoard of gold coins; though there has to be a certain amount of intelligence involved in recognising the significance of the details that one has by chance spotted.
I thinking making observations and then forming theories based on those observations sounds, on the face of it, very reasonable. However the hypotheses one forms determines the questions one asks and so the data/observations one chooses to compile.
If the evidence that I have seen so far leads me to the hypothesis that the Voynich is written in cipher then the questions I ask of the data and the manuscript as a whole are different than if I think it is written in a natural language. If the evidence leads me to the hypothesis that the manuscript is 20th century then the questions that I ask will be different than if I think it much more likely that it is 15th century. One might explore the hypothesis that it is Mexican instead of say Polish and ask different questions on that basis.
I think one explores working as yet unproven, but seemingly plausible, assumptions to see what the consequences of those assumptions might be and whether the consequences fit with the data. A lot of the time I think one asks “what if” questions I.e. what if assumption X is true?, leaving one to explore the consequences to see if that takes one anywhere interesting.
I question the notion that one can work without forming hypotheses. If one chooses to carry out a frequency count of glyphs in the Voynich then one is making the assumption that that might be meaningful and useful information, based on ideas about language patterns, whereas it may not be; nevertheless it seems a very plausible hypothesis that this data might give insight into underlying phenomena.
I think if one acknowledges that oneself like every other researcher has inevitably got some things wrong then one should be open to the possibility of changing one’s mind.
Also, it benefits the research community as a whole if different people explore different hypotheses as it results in a wider examination of the solution space, which would be less likely to happen if people are more cautious. I am not overly fond of the extra cautious approach, as one can easily end up not leaving one’s house, so to speak.
Mark: the big difference is that Alice Kober actually decrypted (all bar the final step, which Ventris got the accolades for) an unknown language in an unknown script. As such, I think Cipher Mysteries readers ought to be aware of her viewpoint, hmmm?
Nick: I certainly think it is worth being aware of her viewpoint and it is good that you made us aware of it.
(There is some suggestion online that there was a bit of a bias towards Kober by the author of the book, but given that I haven’t read the book nor studied the subject in detail I would not venture an opinion on that.)
Nick: Did you read “The Man Who Deciphered Linear B: The Story of Michael Ventris” by Andrew Robinson?
It does sound very plausible that the work of Alice Kober has been sidelined. Nevertheless I think we all gravitate to the idea of there being one hero of a story.
This guy Andrew Robinson also wrote a book about Champollion. I have read a bit in the past about Champollion, but not a lot. I agree it definitely has value looking at other people’s methodological approaches.
Mark: all the previous books were written without having seen Alice Kober’s archived notes. Ventris didn’t give Kober the credit she deserved.
Hi ants. The key to the manuscript is to know the Czech language.
I bet a 10 000 dollars with every scientist.
Nick, you´ve been trying to figure out many years what stars mean, Which have an S- shaped comma.
In the last article I write about it. This is the 70v page.
Big circle : It describes 10 births of Anna Hlohovská. 2 stars drawn together – twins children.
6 women are dressed and 4 are without clothes. That means 6 daughters and 4 sons.
Small circle : Two births Eliška ( Elizabeth of Rosenberg ). Two stars together – birth twins. Eliška gave birth twice and had three sons. Therefore, there are 2 women and 3 sons.
Anna Hlohovská was the mother of Eliška. And the wife of Jan II. Roesnberg.
Everything is written in the Czech language.
And the wife of Jan II. of Rosenberg.
Mark – I take your point. Working to understand a written text is different from iconological analysis which proceeds – if done well – rather differently. I guess if I were to reduce the method to its absolute basics – you look at an image, notice the points at which it does or doesn’t conform to the ‘rules’ for the time and place previously attributed to it and either confirm the attribution and interpretation or take as much time and trouble is needed to correct previous errors. ‘Where and when did this style of drawing, a palette of this range… etc. occur in fact?’ Much less need for hypotheses – there’s so much real information out there. Not so with Voynichese or with Linear B, back then.
Fascinating story! But – not to forget recent machine learning approaches: https://arxiv.org/abs/1906.06718.
Wow, extraordinary !
Not being in the field, I knew nothing about Kober, had never heard of her, and so simply believed that Ventris was the one who made the breakthrough based mainly on hearsay,
In addition, reading over Nick’s suggested approach, I find my approach has massive similarities with Ventris’ initial, and wrong, approach. In my case, after casting about for a likely suspect, I decided on Dr. Arthur Tucker’s meso-American hypothesis. He has unfortunately passed away but for several years was my mentor in the area of meso-American herbology, pointing out and even sending me reference materials and books and answering every question. I must confess I had more of an interest in Aztec herb use than the Voynich but learned much in both areas. His book, “Unraveling the Voynich Codex”, co-authored with Jules Janick, remains a masterpiece of socio-cultural speculative analysis, irrespective of the correctness of the hypothesis.
For some time now, I have entertained the idea of taking a new approach to Voynich, more analytical and less presumptive, but the sheer amount of work for this is frightening. I shall read more on Kober’s approach, as always, with thanks to Nick for the usual highly intriguing post. Her construction of index cards and the other work she did shows the most extraordinary determination and it is my hope that she will get more recognition over time.
As a general point I think when it comes to the question of priority and credit we have to be aware of the fact that it is quite common for 2 people to reach the same conclusions independently.
The most famous example that occurs to me is that of the argument over who invented Calculus. Now Newton’s supporters claimed he did whilst Leibniz’s supporters claimed he did. The modern understanding is, I believe, that they invented Calculus quite independently, though Newton likely invented it first, but published second, if my memory serves me well.
My point is that when the Voynich is deciphered I can see someone saying that their ideas were knowingly plagarised by the decipherer, and in justification they can point to somewhere amongst the pages and pages of writings on the Voynich on the internet to prove it was their idea. Now their idea could have been plagarised by someone who read their thoughts on their blog or Ninja or here or somewhere else. However one shouldn’t assume the decipherer to have read all of the vast amount of Voynich writing on the internet and it is perfectly possible and very probable that the decipherer independenty had the same idea as someone else. Therefore it is perfectly possible that someone fails to give someone else credit simply, because they were unaware of their work. When the Voynich is deciphered I wouldn’t be shocked if there was a bloodbath over the subject as to who was the originator of an idea and whether something was plagarised. The debate over how the credit should be shared out could be greater than the prior debate over what the solution is.
It’s very disappointing that when describing achievements of one researcher one has to belittle the role of another one. But at least Ventris’ name was somehow typed in – I have read texts about decoding Linear B where only Kobler was mentioned:)
Yes, it maybe that Ventris just slightly upgraded Kobler’s theory. But then we should agree that Einstein just reduced special relativity to two simple postulates to give a physical ground to long-known Lorentz-Poincare transformation. And a striker just nets the ball ending a huge work done by the whole team.
By the way have you read that Einstein just couldn’t contribute to physics because of his weak abilities in maths and that all mathematical work for him was done by his assistant or even by wife Mileva? Or was it because of his slight contribution to physics we consider him to be weak in maths? Hm… always forget the right direction of this implication…:(
Yes, technically it can be tiny detail (and it often is) which allows to make a huge leap, but the thing is that some people posses a talent to add this detail while others are stuck (maybe temporary, but still).
Returning to Voynich it’s possible that someone will produce twenty seven foolish theories but the twenty eighth will be the one. That is absolute disaster from educational point of view, but to keep collecting stats and data without any risky moves hoping that some day the solution will magically appear by itself is the worst alternative.
Nick, for deciphering a script it is necessary to analyze the way the script works. For doing so it is required to develop a set of methods and tools that fit to the peculiarities of that script. But the most important point is the willingness to challenge every idea you have. This is what is necessary to decipher any unknown script and this is what Alice Kober and Michael Ventris have done.
Torsten: what I personally took from reading about Alice Kober’s sustained assault on Linear B was the huge importance of analyzing the script on its own terms, not in terms of any hypothesis concerning the language (or indeed the possible lack of language) beneath the script.
It’s a hard path – and Kober’s life shows exactly how hard it is – but perhaps it will prove to be the only genuinely productive path.
farmerjohn: Margalit Fox’s book is well worth a read (and it’s not expensive second hand), I think you’d enjoy it.
I hear exactly what you’re saying about how some people have attempted to rewrite Einstein’s life and works, but Kober’s story really isn’t that kind of thing. Read the book and make up your own mind. 🙂
Nick, to analyze the scripts on its own terms is the way every decipherment of an unknown script was done in the past. The reason for that is that only by understanding the writing system you can hope to find clues pointing to the language behind the script. For instance Kober noticed that common roots in Linear B had three different suffixes. This fact told her that Linear B is an inflected language and was demonstrating the existence of three grammatical cases. For identifying the language this are invaluable clues.
Torsten: I completely agree with you when you say “But the most important point is the willingness to challenge every idea you have.” It is very easy to get trapped in one mindset and that can fatally hamper one’s progress.
Mark – if two people should *ever* claim to have reached the same conclusion about Voynichese independently it will be a very good thing. Should it happen, we can leave it to people with real skills (independent peer review level) to decide if both did so honestly.
btw Citing precedents isn’t normally about squabbling over credit. It’s about making clear that one is neither (1) woefully ignorant or (2) irrationally greedy.
An example of (1) cropped recently here. Nick had to explain to some chap excited about his ‘discovery’ that it was pretty much the same as Currier said more than half a century ago. Had the chap done his preliminary studies, and read that precedent study, he might have save himself and others wasting time, and been further saved public embarrassment.
Of course, the reason Currier’s work is so well known is that before the early 2000s, the same standards normal everywhere else also applied within Voynich research.
d’Imperio is pretty straight-up about sources and precedents. So were most members of the first (Reeds) mailing list – and Nick is among those who understand that in service to the study, and to readers, it is important to mark the division between one’s own thoughts and information gained from other people.
“His book, “Unraveling the Voynich Codex”, co-authored with Jules Janick, remains a masterpiece of socio-cultural speculative analysis, irrespective of the correctness of the hypothesis.”
I’m sorry, I don’t usually come right out and say things like this, but this book is an example of hasty, theory-driven assumptions, poor iconographical analysis, internally inconsistent attempts at proposed relationships, and woefully inadequate knowledge of medieval history and scribal conventions. It is even questionable in terms of some of the botanical assertions.
It is the anti-example of how to do research.
Torsten: oddly enough, the only thing that stopped Kober from making the final step was a single preconceived idea about those triplets. Fox p.266: “But recognizing the triplets as having derivational rather than grammatical endings is what let Ventris surpass the ground she had gained.”
You should read the book, it’s well worth it. 🙂
Nick, I have already set the book on my reading list.
I wonder if you want to read my book about the Phaistos disc from 2008. The book also contains a chapter about the theory of decipherment of forgotten scripts including the decipherment of Linear B.
Torsten: your book has been on my reading list for some time, but I haven’t really been thinking about the Phaistos Disc much. Perhaps it’s time to take a look. 🙂
There is a solution to Linear A that is described in the following video:
“Breakthrough Decipherment of Minoan Linear A and Cretan Hieroglyphs”
https://www.youtube.com/watch?v=PiLyN9T2stY
https://www.smithsonianmag.com/history/have-scholars-finally-deciphered-a-mysterious-ancient-script-180980497/
This refers to Linear Elamite.