A copy of Paul de Saint-Hilaire’s (1973) “La Belgique Mysterieuse” arrived in the post today. I saw it mentioned here, and ordered it because it seemed to be the only book out there (apart from Rudy Cambier’s Knights Templar / Nostradamus stuff) where Moustier Church’s strange cryptograms are discussed at all.

It’s a handy little thing, that divides its mysterious subject matter into:
(1) Megalithic Belgium
(2) Templar Belgium
(3) Lost Treasure Belgium
(4) Alchemical Belgium
(5) Fantastic(al) Belgium
(6) Underground Belgium, and – of course –
(7) Secret Belgium.

Each chapter finishes up with a suggested tour around the Belgian countryside to take in the major landmarks it discussed. So… what does our man say about Moustier, then?

* On p.60, Saint-Hilaire says that the cryptograms may relate to a treasure hidden in the 19th century. (But without giving any sources or references – annoyingly, there is no bibliography).
* On p.119, he notes that the cryptograms are laid out in the same way as you would expect to see “The Tables of the Law” (i.e. the Ten Commandments), but that it has so far proved impossible to find the key.
* On p.129, he speculates that examining some windows of the church and a small chapel in the village might prove useful in decrypting the inscriptions. (But, again, without giving any sources or references).

And that’s all Saint-Hilaire says about Moustier. It’s not much, true: but it is what it is.

* * * * * * *

Having mused on the Moustier enigma for the last few weeks, my own conclusions are:
* I don’t believe that it was carved before 1800
* I don’t believe that it uses a simple substitution cipher
* I don’t believe that it uses a transposition cipher
* I do believe that it includes copying mistakes
* I do believe that it was copied from a written version, not from a carved text
* I do believe that the written version was probably devised for the rebuilding of the Church
* I do believe that the resemblance to The Tables of the Law will prove to be no coincidence

Put all that together, and I suspect that the cipher that was used was very probably what is generally known as the Vigenère cipher (even though it was actually first invented by Bellaso, but we’ll let that pass).

The immediate question is, of course, is whether this is historically plausible for the proposed date? I’m sure it is – in fact, at that very time, it was renowned as le chiffre indéchiffrable, i.e. ‘the unbreakable cipher’.

I also have here a fascinating (if occasionally meandering) book by Ole Immanuel Franksen called “Mr Babbage’s Secret”, which reveals that Charles Babbage had quietly worked out the basic principles of how to crack Vigenère ciphertexts as early as 1846. This was later independently worked out by a Frenchman called de Viaris in 1888, who published his results in the “Génie Civil”, but without attracting much attention.

Because the Vigenère cipher was particularly popular in France, I predict that the plaintext will turn out to be in French: and moreover, I suspect that the key word will turn out to be something to do with the Ten Commandments (or perhaps Christianity in general), or perhaps even the name of something featured in a church window (as Saint-Hilaire speculated).

Anyway, that’s pretty much as good a set of constraints as my historical sleuthing has been able to generate, and I suspect there’s no more useful information out there we can get our hands on. So now it’s probably time to move to phase #2: checking the transcription and doing a bit of cryptanalysis. If it’s a Vig, can we work out what its key-length is? Hopefully we shall see! 🙂

22 thoughts on “Paul de Saint-Hilaire’s notes on Moustier Church…

  1. bdid1dr on July 13, 2013 at 1:03 am said:

    Nick, I’ll root around in my brainpan (where I keep vague memories of various bits and pieces of writing which catch my eye. If I am successful, I’ll get back to you — but only if you give me the “go-ahead”. Cya in a couple of days. 🙂

  2. bdid1dr on July 13, 2013 at 2:46 pm said:

    Nick, correct me if I am not recalling details clearly: I took a look at your earlier Moustier discussion page and photos. Did not the tablets appear in two separate sanctuaries — one being the “Lady Chapel”? If so, could there be clues, there, as an explanation of a possible meaning hidden in the tablets?

  3. bdid1dr: could be. It’s certainly a place I’d like to go and examine closely all over, who knows what’s there to find?

  4. SirHubert on July 18, 2013 at 5:22 pm said:

    Not very much to go on for finding a Vigenere keyword.

    Two pairs of trigrams appear on the St Martin altar: BΓP, starting at letter 5 and letter 64, and RAL/RΛL (assuming that A and Λ are the same), starting at letter 30 and letter 47. Can’t find a common factor for those.

    I’ve tried the positions of the letter pairs as well. On the Virgin’s altar PN appears at letters 40-41 and 80-81, and ΛL at letters 32-33 and 67-38. That looks a little bit like spacing divisible by 5, but it’s pretty tenuous and other bigrams don’t seem to follow the same pattern. And of course ΛL appears in the RΛL trigrams which are spaced 17 letter apart, and 17 isn’t divisible by anything useful…

    I can see why you would think the VIgenere is historically and geographically appropriate, but it somehow seems not quite right in this context. Apart from substitution ciphers, what other possibilities do we have at this period?

  5. SirHubert: I know all *that*… but I suspect with such a short text, a Vigenere would indeed typically produce stats of that general order, including the one-possibly-two 3-gram(s). As I recall, the stats only go really flat if you have a longish key-length and non-repetitive key.

    Circa 1838, people thought it was (literally) “indechiffrable”, so why bother using a really long key?

    As for the non-multiple repeat length (59, is it?) I have to say that I suspect copying errors may have messed things up – not hugely, but enough to throw things out a little.

    Hence today’s best guess is that the gap should be a more multiple-friendly 60 and so the key-length will probably be 6, 10 or 12. 🙂

  6. Putting all the pieces together, I predict the key-length is 6, and that a character was omitted between #30 and #47. 🙂

  7. SirHubert on July 18, 2013 at 10:09 pm said:

    Nick: well, maybe. You’re clearly pretty sure this is going to be a Vigenere, even though the stats as they stand don’t really support this. So you’re now supposing a textual error which would make the inscription fit your preconceived hypothesis. Isn’t that what you’ve just been criticizing Voynich researchers for doing? But it’s hardly a huge correction to make, and MARTIN and MARIE/MARIA would give nice six- and five-letter keywords for their eponymous altars.

    Personally, I’m not sure that the relative crudeness of the engraving and the peculiar letter-forms match what was then state-of-the-art encryption. And if the ciphertext was newly written for the altars, wouldn’t it have been terribly easy for the writer to ask the mason to pencil the inscription on the white marble so it could be checked before engraving?

    And what’s with all the strange unbarred letter forms? It’s all Greek to me 😉

  8. Not necessarily *Greek* but maybe the Moustier tablet-maker’s attempt to re-write or reproduce the “Egyptian” or”Hebrew” words which Moses (?) was commanded by YahWeh to entablature? Exodus 34 et seq?

  9. Nick, I am pretty sure, that the Moustier text has been written in the Walon language (ca. Charleroi) and that it is not a code. Only the first letters of words are indicated like on the crucifix text INRI (or JNRI), which stands for IESVS NAZARENVS REX IVDAEORVM. Furthermore the Q sign should be interpreted as O (which is missing in all four texts).

    My transcription is:

    J N L K B F P R
    V M G H W H”
    Q L S B N F H*
    M G E K H V R
    A L R N F S X V
    P F V B L P M R
    R A E G K T D
    B N D F J V R W
    L U B F P N I D
    C E T R A Q M

    L F E G K R V Q
    Y P Z H N R L B D
    M F A N V D E
    N A P V J H M A
    L F N A E B K P
    N C L X B P D W
    R N E C H Z R P
    M D X R A P L N
    H F A L D N X W
    E N L V N D A P N

    in which M, W and the ligatures use double space. Normal length of line: 8 characters. I think the text is a religious text, maybe taken from the Life of St. Martin by Sulpitius Severus. I have not yet identified the text, but this may be a step forward.

  10. SirHubert on September 20, 2013 at 11:10 pm said:

    Menno: again, perhaps. There are still three letter-forms to explain which don’t resemble elements of the standard Latin alphabet: the ‘gamma’, ‘lambda’ and ‘square bracket’. I know very little about Walon/Walloon as a language, but as far as I can see it just uses the standard Latin characters plus a few accents on certain vowels.

    The two altars are dedicated to different individuals: St Martin and the Virgin respectively. I don’t know that anyone has determined whether or how the two inscriptions are linked, or even whether they were produced by an identical form of encryption (or abbreviation).

    I agree with you that, on the face of it, one might expect the text(s) to be religious. And, as has been suggested, they might relate to St Martin and/or the Virgin (as dedicatees of the altars) or to the Ten Commandments (given that there are ten lines on each). Kind of obvious, but a fair guess, surely? What do the Ten Commandments look like in Walloon?

    Alternatively, if Nick is right and the texts were produced by a Vigenere cipher, can we not guess plausible keywords (‘Maria’, ‘Martin’ etc) and try applying those to the ciphertexts? Unless this particular Vigenere was enciphered using a grid made up of scrambled alphabets instead of the standard A-Zs (which would be really unfriendly)…

  11. SirHubert, the three unbarred letters are F, E, A. No doubt about that. Similar unbarred letters you may find on old and new gravestones. The background is the risk to spoil the masonry, so they left them out. Yet everybody could see the meaning of the unbarred letters. Even in modern times unbarred letters are used in logo’s. You wouldn’t even notice.

    The two double column plates are likely to come from an altar triptych. The middle altar piece is missing (or can be found somewhere else) with a two-column plate left and right. The middle altar piece could give e.g. a picture of St. Martin on a horse cutting his cloak for a beggar. I have not yet found an indication, which of the column pieces stood left or right. The dedication to two individuals (St. Martin and H. Mary) is certainly to be dated later, when the altar pieces were separated.

    In Roman Catholic churches it was common practice to represent words by its first character : INRI, RIP, IHS, etc. The longest in the Roman acronyms and abbreviations is : Q.D.E.R.F.P.D.E.R.I.C.= Quid De Ea Re Fieri Placeret, De Ea Re Ita Censuerunt. Viginere or not, I do not think we deal with a cipher in the Moustier text. What would be the purpose to encrypt a religious text, which is commonly available ?

    I did not find out about the Walon language through the Ten Commandments (too short text), but by comparing the Pater Noster in Latin, French, German, Dutch, Flemish, etc. There is an excellent German Wikipedia site, which gives the Pater Noster in all languages. It is called: Sprachvergleich anhand des Vaterunsers. It includes the Walon language.

    Why Walon ? The Latin script does not use J, K, U, W and Y, Z only for Greek words. The European script uses X in Roman numerals, and Y in Greek words. The Walon script uses both J, K, U, W and Y, X. Yet, the Railway station of Moustier is on the Charleroi-Namen line in Walonia. I wouldn’t call this coincidence.

    My guess that the text comes from the Life of St. Martin by Sulpitius Severus (c.363-c.425) is not just because of the name of St. Martins church, but also because of the publications about the Life of St. Martin mid 19th c., which seem to comply with the date of building the church of St. Martin.

    Menno Knul

  12. SirHubert on September 21, 2013 at 6:20 pm said:

    Dear Menno,

    It’s certainly possible that the ‘Greek’ letter forms are just epigraphic variants of A, F and E. But we have at least two, and possibly three, properly-engraved E’s on the Virgin’s altar, as well as at least one properly-barred A on St Martin’s. And the square bracket character might be an E, but I did wonder if it could also be an I, which is otherwise not found at all (which seemed strange).

    Not being Catholic, I’m not sure whether one would expect a vernacular inscription (as opposed to Latin) to be engraved on an altar at this period?

    I don’t dispute that Walon is a possible candidate for language and take your point about the non-standard letters. Equally, a Vigenere table could easily use all 26 letters of the modern ‘Latin’ alphabet – after all, these are all found in modern French. (And, for what it’s worth, the author of the ancient alphabetic hymn ‘Apparebit repentina’ used all 24 of these, only combining I/J and U/V.)

    I’m far from convinced that the text is a Vigenere cipher, but neither am I sure that I believe in a text of 80 characters length composed entirely of first letters. It would have to be an extremely familiar text for anyone to be able to read it correctly. The point of enciphering a standard religious text, I suggest, would be to make a reader look at it and think about it more carefully than they otherwise might. I think I once read that some Roman tombstones employed anagrams or simple ciphers so that passers-by would take more interest in the inscriptions.

    I like your idea of the two panels having originally formed the wings of a triptych. That being so, the inscription would presumably come from Basecles rather than being newly made for Moustier?

  13. Dear SirHubert, On grave stones it happens as well that some letters have a bar and some do not. Probably the quality and structure of the stone is responsible for the decision to take the risk or not. By the way you find the letter I in the line LUBFPNID.

    The main question is, if the Moustier texts have been coded deliberately to hide the contents of the message. I have not found any reason to do so. In my view the texts have mainly an ornamental purpose. There is a good possibility, that the altar contained a shrine with remnants of St. Martin. In that case the altar stones would be covererd with an embroidered cloth. We don’t know yet.

    The Basecles St. Martin church has a wooden presentation of St. Martin: On y voit dans la niche centrale une statue équestre de saint Martin partageant son manteau avec un pauvre. Elle est en bois polychromé, du XVIIe siècle. The Neoclassical style of the church could have played a role in the St. Martin’s texts. I have not yet looked at the churches’ history.

    Menno Knul

  14. Additionally to my former remarks.

    The Moustier case shows the importance to establish first the script and language of a given text before deciding, if a text is encrypted or not, especially when the environment of a text does not show a need to encrypt a message. In the Moustier case the religious context (whether prayer, psalm, bible text or otherwise) there is no need to encrypt the text because the same text is publicly available in a natural language as well.

    I think the same applies to the VMS. Herbaria were commonly available in natural language, whether in Latin, Italian, German, Spanish or Arabic. So there would be no need to encrypt the text. The same applies to the astronomical and astrological pages.

    Here too our first task should be to identify the script and language before deciding if the VMS is encrypted or not. An indication could be that the VMS contains similar pages (e.g. f71v red and white aries) with a different ‘encryption’, which could show the same (translated) text in different languages, e.g. Spanish and Italian, but in the same unknown script.

  15. SirHubert on September 24, 2013 at 8:50 am said:

    Dear Menno,

    The VMS has been studied pretty thoroughly to determine whether a natural language underlies the text. That this topic is still being debated so heatedly tells its own story, I think!

    The question of where and when the altars and their inscriptions were carved is probably better answered through church records. I do quite like your suggestion that they may have formed the wings of a triptych, but if each contains half of one long inscription I don’t see why they’d be separated in this way. Unless no one could understand them, of course…but then why put up altars with texts you didn’t understand?

    Oh well. Back to playing with a text I’ve probably mis-transcribed, containing a language I may not be able to read, enciphered by a system I don’t understand.

  16. SirHubert
    May I ask if you’ve published any of your VMS work online?

  17. Dear SirHubert,

    I think basically the text (which ?) has been divided to make four equal columns including wide letters (M,W) and ligatures.

    Menno Knul

  18. SirHubert on September 24, 2013 at 4:36 pm said:

    Diane – that’s something I’ll consider if I ever come up with anything which might conceivably be worth publishing!

  19. SirHubert
    speaking of Greek, how’s your cursive?

  20. SirHubert on October 4, 2013 at 4:50 pm said:

    I’ve come across it occasionally but I can’t claim any real expertise. More promising than Pahlawi, if you’re asking what I think you’re asking? What can you tell me about Hephthalite imagery?

  21. I very much like the idea of the Hepthalites as an historical thesis for the Vms

    If you rely on art histories, you could make an argument from that basis, I should think, especially in the far west after the fourth century invasions.

    I wouldn’t ever attempt to argue a consistent ‘Hepthalite art’. I don’t think it’s feasible given the range, time, and uncertainties – even about the terms found in primary sources.

    For the Voynich ms, there’s one immediate problem. Most of its nymphs, apart from the calendar section which is ‘odd’ in lots of ways – have long heads. Short heads are one of the few elements considered characteristically Hepthalite.

    Perhaps I can illustrate by quoting. Comments in square brackets mine.

    “The fourth vessel.. represents the best known specimen of Hephthalite art known up to date. ..the body of the bowl is decorated with six dancers in Indian costume with Iranian ribbons and Hephalite-short heads. They are standing in .. arcades whose columns form vegetable convolutes in the style of Taq-I Bustan, carrying two [characteristically Persian] bull figures as capitals.. capitals recall the ivories of [Greco-Indian] Begram, the capitals of the [early Buddhist] Bharut pillars or … other Indian cave sites. The spandrels hold en face two-winged and crouched women.. which might go back to the [Indo-Greek] gandharva type..

    So is a bowl like that an Indo-Greco-Persian work produced for a Hepthalite ruler, or is it to be considered a uniquely Hepthalite work? Art is often defined for convenience by eras rather than by internal reference e.g. the ‘roman’ glass found in Begram probably never saw a Roman in its life. 🙂

    quote from
    B. Brentjes, ‘The Hoard of Hephthalite Silver Vessels Found near Samarkand’, East and West , Vol. 21, No. 1/2 (March-June 1971), pp. 77-78.

    PS the ‘Norman’ style helmets occur so widely – not much useful there. But I’ve sometimes wondered about whether the Normans were conscious of inner Asian origins. Frederick II claimed and was granted by the Pope, ‘the lands beyond Antioch’. And Norman costume looks so Sogdian!

  22. For others reading this and wanting an overview, the article ‘Huns’ in Encyclopedia Iranica is more sober than any of the wiki articles I’ve seen so fr

    . Assertions of an eastern ‘Hephthalite empire’ are unusual.

    For the Chinese terms so often mentioned, Hill’s notes my be helpful.

    John E. Hill,
    Notes to
    The Western Regions according to the Hou Hanshu.

    I have a copy of the 2003 (2nd) edition, but one more recent may be online.

Leave a Reply

Your email address will not be published. Required fields are marked *

Post navigation