+- The Voynich Ninja (https://www.voynich.ninja)
+-- Forum: Voynich Research (https://www.voynich.ninja/forum-27.html)
+--- Forum: Analysis of the text (https://www.voynich.ninja/forum-41.html)
+--- Thread: The Voynich Manuscript is a cipher because...? (/thread-407.html)
RE: The Voynich Manuscript is a cipher because...? - Sam G - 20-02-2016
One of the most important facts about the text is that the words are "pronounceable", in that there's a mapping of the individual glyphs onto the Roman alphabet that makes the words basically pronounceable in English. Obviously this fact is not true for many/most types of ciphertexts. The fact that this mapping preserves letters esssentially identical to "a", "e", and "o" as vowels shows that this is not coincidental. We could also probably make the words even more pronounceable if we allowed the use of a phoneme inventory of a language other than that of English. It seems impossible that this could be completely accidental. So already we have moved from "language, cipher, or gibberish" to "language, ciphertext that resembles unencrypted text, or gibberish that resembles unencrypted text".
Next we have the low entropy of the text. This is perfectly consistent with unencrypted natural language text in an alphabetic script, as shown by the fact that Japanese and Hawaiian (and probably countless other languages that haven't been checked) have a lower entropy than the VMS text, whereas the only type of cipher that produces such low entropy text is verbose cipher, and verbose cipher can be ruled about because the words are simply not long enough, and there aren't enough common two-word sequences to assume that the words have been broken up into multiple smaller words.
So, already we're looking at language-like text that cannot have been produced by any known cipher mechanism. As of 2016, nobody has ever developed a cipher capable of replicating even a fraction of the VMS text's properties, despite the enormous advances in cryptography that have been made since the VMS was produced, and the fact that some of the best cryptographers in the world have worked specifically on the VMS.
Other issues:
The <p> and <f> gallows in initial lines of paragraphs are often cited as something non-linguistic, but actually many old Latin manuscripts have these kinds of embellishments. Do they ever show up in ciphertexts? If not, then I'd count them as another point in favor of unencrypted natural language text.
The fact that the alphabet is "invented" is cited as evidence favoring cipher, but adapting existing scripts to write new languages is how most scripts in use in the world came about, and I'd argue that the VMS script is no different.
One of the most remarkable aspects of the VMS script is that it is designed to highlight and emphasize the structure of the text. This is completely contrary to the purpose of cryptography, which is to conceal any structure in the ciphertext in order to make it more difficult to crack. However, creating such a script is perfectly logical to do if you are designing one for a natural language.
Structural properties of the text aside, it's hard to see how a ciphertext resembling the VMS would ever be put to use. How would someone read it?
For instance, could someone who understands the "system" involved here read the VMS "as is", or would he need to work out the plaintext on scratch paper (errr.... guess that would be "scratch parchment")? It seems that only simple types of ciphers (substitution etc.) would be readable without writing out the plaintext, and these can all basically be ruled out in the case of the VMS. So we have our potential reader of the VMS copying out the plaintext by hand.
But what about the illustrations and circular diagrams? Does he copy these out as well? Some of them would seem hard to use without the associated text being deciphered too. What would he do with all the labels? Would he write the unenciphered text directly on the VMS? Or maybe "tape" the deciphered text on in some way (did they have anything like tape back then?). It seems hard to see how he would do it without copying all the illustrations. And then does he add all the colors as well? This is really sounding like quite a lot of work to use this thing. Is there any precedent for this?
And now that our VMS reader has gone through all the trouble to decipher the manuscript, and has the unencrypted plaintext sitting in front of him, we have to ask: what was the point of encrypting it in the first place? To protect its secrets while it was in transit? Isn't this a lot of trouble to go through to protect it while it was in transit? And what about the VMS is even worth keeping secret? A bunch of information about herbs? A zodiac? Does that make any sense? Is there a known example of an encrypted herbal or an encrypted zodiac?
Some people say it was encrypted to protect its creator from the church. Apparently this is the church that hates information about herbs, but has no problem with naked women swimming in pools of green fluid.
It's hard to see a sensible motive here. I guess you could argue that there doesn't necessarily have to be one, and that the creator of the VMS could have been foolish or insane (despite having apparently been one of the greatest cryptographers of all time). But when you combine the fact that the properties of the text itself are incompatible with any known type of cipher with the fact that many other subtle aspects of the VMS seem incompatible with a cryptological intention, it seems more reasonable to conclude that we're not looking at a ciphertext here. Especially since I don't see any convincing positive evidence for a ciphertext being presented.
RE: The Voynich Manuscript is a cipher because...? - Emma May Smith - 20-02-2016
(20-02-2016, 10:52 AM)ReneZ Wrote: You are not allowed to view links. Register or Login to view.(20-02-2016, 05:44 AM)-Job- Wrote: You are not allowed to view links. Register or Login to view.De rhein een cepain faals mayneli oen the pelain.
In fact, this type of alternative and inconsistent spelling (though maybe not quite as adventurous) should probably be expected.
I wouldn't call it part of a cipher....
When I occasionally try to read German texts from 15th-16th Century books and MSs, they come across as something like the above, compared to modern German.
Would this not invalidate the suggestion that the lack of word order is unlinguistic?
RE: The Voynich Manuscript is a cipher because...? - -Job- - 24-02-2016
(20-02-2016, 10:52 AM)ReneZ Wrote: You are not allowed to view links. Register or Login to view.(20-02-2016, 05:44 AM)-Job- Wrote: You are not allowed to view links. Register or Login to view.De rhein een cepain faals mayneli oen the pelain.
In fact, this type of alternative and inconsistent spelling (though maybe not quite as adventurous) should probably be expected.
I wouldn't call it part of a cipher....
When I occasionally try to read German texts from 15th-16th Century books and MSs, they come across as something like the above, compared to modern German.
Are there any sample transcriptions of texts with irregular spelling and/or abbreviations available for analysis? They would be quite valuable.
Spelling and abbreviations are plausible justifications for some of the word variability in the VM, but i'm skeptical.
RE: The Voynich Manuscript is a cipher because...? - -Job- - 24-02-2016
(20-02-2016, 06:07 PM)Sam G Wrote: You are not allowed to view links. Register or Login to view.It's hard to see a sensible motive here. I guess you could argue that there doesn't necessarily have to be one, and that the creator of the VMS could have been foolish or insane (despite having apparently been one of the greatest cryptographers of all time). But when you combine the fact that the properties of the text itself are incompatible with any known type of cipher with the fact that many other subtle aspects of the VMS seem incompatible with a cryptological intention, it seems more reasonable to conclude that we're not looking at a ciphertext here. Especially since I don't see any convincing positive evidence for a ciphertext being presented.
You're mostly arguing for the general improbability of a cipher, in several cases using the logical fallacy of argument from incredulity.
Most VM theories may be depicted as improbable. For example, what's the probability that the VM contains an unreadable natural language which exhibits no clear association with any other known language?
Your premise that low entropy is only compatible with a verbose cipher is false, AFAIK. For example, null characters can also lower the character entropy depending on how they're employed.
Also, relating the VM's low character entropy with Japanese or Hawaiian without accounting for its other properties is comparable to classifying the VM as a verbose cipher without accounting for the word lengths.
Is there a language or cipher that matches all of Voynichese's features? I think that should be the focus of this discussion.
RE: The Voynich Manuscript is a cipher because...? - ReneZ - 24-02-2016
Assuming that the author had a plain text, either in his head or in writing before him, and he somehow converted this to the Voynich MS text...
Then the low entropy either was already there in the plain text, or it is a result of his operation.
Looking at this operation, the random insertion of null characters is more likely to increase entropy, unless it wasn't really random but followed some pattern, in which case it isn't fundamentally different from a verbose cipher.
Compression algorithms (like the well-known 'zip') specifically increase the entropy, and the uncompression decreases it.
To decode the Voynich MS text, one is then looking for something similar to compression.
Removal of nulls would be a form of compression, and combining characters another.
Things I have seen in some translation attempts, e.g. suggestions that the gallows characters represent the Nahuatl character pair 'tl', are going into the wrong direction.....
Another suggestion: that the text is vowelless, and vowels need to be inserted, is equally going in the wrong direction. Such a further reduction of entropy would need to be balanced by 'something else' that stronly increases it.
RE: The Voynich Manuscript is a cipher because...? - Sam G - 24-02-2016
(24-02-2016, 09:18 AM)-Job- Wrote: You are not allowed to view links. Register or Login to view.(20-02-2016, 06:07 PM)Sam G Wrote: You are not allowed to view links. Register or Login to view.It's hard to see a sensible motive here. I guess you could argue that there doesn't necessarily have to be one, and that the creator of the VMS could have been foolish or insane (despite having apparently been one of the greatest cryptographers of all time). But when you combine the fact that the properties of the text itself are incompatible with any known type of cipher with the fact that many other subtle aspects of the VMS seem incompatible with a cryptological intention, it seems more reasonable to conclude that we're not looking at a ciphertext here. Especially since I don't see any convincing positive evidence for a ciphertext being presented.
You're mostly arguing for the general improbability of a cipher, in several cases using the logical fallacy of argument from incredulity.
I am not using argument from incredulity. Attempting to understand the motives of the VMS creator from the decisions that he made is not fallacious, and it would seem to me that considering how a document such as the VMS is intended to be used would be pretty important to understanding what it is and why/how it was made. Actually I think it's a fairly standard thing to do with any document or even any artifact. I also note that you did not actually address any of the points I made.
Quote:Most VM theories may be depicted as improbable. For example, what's the probability that the VM contains an unreadable natural language which exhibits no clear association with any other known language?
What I'm trying to do is figure out what the VMS is by looking at the evidence, rather than deciding what it is a priori. I obviously don't consider the idea that a medieval European document could be written in an otherwise unattested and presumably non-European language to be so unlikely as to rule it out a priori, but then I tend to assume that there's much about history that we don't know - you might feel differently.
This is the major problem with the theorizing around the VMS as far as I'm concerned - the tacit assumption that the VMS must be historically unimportant and must not contain anything fundamentally new or significant, which leads to the view that it cannot be an unencrypted natural language text no matter what the evidence suggests (because this would make the VMS historically important for a number of reasons), which then leads to all the unworkable cryptological theories that are in complete contradiction to the available evidence. I might write a separate post on this topic.
Quote:Your premise that low entropy is only compatible with a verbose cipher is false, AFAIK. For example, null characters can also lower the character entropy depending on how they're employed.
This could be considered a form of verbose cipher, and you'd still have the same problem: short word lengths.
I think lossy substitution ciphers (where you map multiple plaintext letters to the same ciphertext letter) can also lower entropy - if I'm not mistaken, Jim Gillogly points that out as the only other entropy-lowering cipher he could come up with somewhere in the list archives. But again, it can be ruled out in the case of the VMS, at least for European languages.
Quote:Also, relating the VM's low character entropy with Japanese or Hawaiian without accounting for its other properties is comparable to classifying the VM as a verbose cipher without accounting for the word lengths.
Doesn't really matter as far as refuting the cipher theory is concerned. If there's some reason to think that a natural language with low entropy could not possibly possess some property found in the VMS text, it would not really help the cipher theory, but only undermine the natural language theory, and support the gibberish theory - but I don't see that there is any such reason.
Quote:Is there a language or cipher that matches all of Voynichese's features? I think that should be the focus of this discussion.
Not a known language or a known cipher, of course - that seems pretty clear, though I guess there are still plenty of languages in the world that haven't been checked. In theory it could be mostly the same as a language that's still in use by a small population, though I doubt this. So the question is really whether it's an unknown language, or a cipher mechanism that remains unknown and undemonstrated as of 2016, or a form of gibberish produced by a mechanism that remains unknown and undemonstrated as of 2016.
RE: The Voynich Manuscript is a cipher because...? - MarcoP - 24-02-2016
(20-02-2016, 06:07 PM)Sam G Wrote: You are not allowed to view links. Register or Login to view.One of the most remarkable aspects of the VMS script is that it is designed to highlight and emphasize the structure of the text. This is completely contrary to the purpose of cryptography, which is to conceal any structure in the ciphertext in order to make it more difficult to crack.
....
And what about the VMS is even worth keeping secret? A bunch of information about herbs? A zodiac? Does that make any sense? Is there a known example of an encrypted herbal or an encrypted zodiac?
Hello Sam, I am not an expert in the field of medieval cryptography, but I would like to share with you some evidence that you might consider. About the structure of the text, are you familiar with Fontana's encrypted manuscripts (e.g. Bellicorum instrumentorum liber cum figuris - You are not allowed to view links. Register or Login to view., Venice, 1420-1430)?
The structure of these manuscripts is often similar to what we observe in the Voynich ms: a large illustration with an encrypted text. The main difference is the presence of titles or short summaries in non-encrypted Latin.
About herbals:
A few years ago, Nick Pelling mentioned a paper that is now available You are not allowed to view links. Register or Login to view.:
"The Cipher Alphabet of John de Foxton's Liber Cosmographiæ" by John Block Friedman. Foxton's book was partially encrypted: the encrypted words are mainly about medicine (a subject that was certainly related to herbals in medieval culture).
The paper also mentions another manuscript: Bodleian Digby MS. 69, c. 1300, a collection of experiments of various kinds in which a single group of recipes —of special value — is encoded, and introduced by the phrase « Expérimenta quadam in quibus occurrunt multa verba scripta in forma occulta omissis vocabulis ». You are not allowed to view links. Register or Login to view. also includes a herbal ("Macri poema de virtutibus herbarum").
In 1200 ca, Hildegard Von Bingen created a secret language ("Lingua Ignota"), written in a newly created alphabet. Hildegard was a physician and herbalist. 181 of the 1011 words in the Lingua Ignota are You are not allowed to view links. Register or Login to view..
Finally, I have been working on the encrypted marginalia of BNF Latin 7272 "Tractatus de sphaera" by Andalius de Nigro, 1330 ca. (I read about the manuscript on You are not allowed to view links. Register or Login to view.). I have posted my tentative decryption and translation on You are not allowed to view links. Register or Login to view..
All the encrypted annotations are about astrology:
* the first annotations are about the Zodiac signs (and mention the name of an angel associated with each zodiac sign)
* one is about the planetary god Saturn
* the largest set is about the Mansions of the Moon (they associate a planet to each mansion and add details about its specific features: what should or should not be done when the Moon is in that position)
It seems to me that, even if the surviving evidence of Medieval cryptography is quite limited, it is compatible with what we observe in the Voynich manuscript.
RE: The Voynich Manuscript is a cipher because...? - -Job- - 24-02-2016
(24-02-2016, 03:20 PM)Sam G Wrote: You are not allowed to view links. Register or Login to view.I am not using argument from incredulity. Attempting to understand the motives of the VMS creator from the decisions that he made is not fallacious, and it would seem to me that considering how a document such as the VMS is intended to be used would be pretty important to understanding what it is and why/how it was made. Actually I think it's a fairly standard thing to do with any document or even any artifact. I also note that you did not actually address any of the points I made.
I interpret the second half of your original post as arguing that the cipher hypothesis is improbable because you can't relate to how or why it might have been used. I understand your reasoning in most cases but it's not a strong argument and not something that can be discussed without introducing significant speculation.
Your point about verbose cipher and word lengths i agree with and won't dispute (nor is it my job to), though it's worth mentioning that, if you accept the possibility of a cipher, then spaces may play a different a role.
The low entropy of the text is not exclusive to a verbose cipher, or at least i don't see why it should be. For example, a code book approach using roman numerals would probably have low second-order character entropy as well, even though it is the opposite of a verbose cipher.
RE: The Voynich Manuscript is a cipher because...? - Emma May Smith - 24-02-2016
(24-02-2016, 11:28 AM)ReneZ Wrote: You are not allowed to view links. Register or Login to view.Assuming that the author had a plain text, either in his head or in writing before him, and he somehow converted this to the Voynich MS text...
Then the low entropy either was already there in the plain text, or it is a result of his operation.
Looking at this operation, the random insertion of null characters is more likely to increase entropy, unless it wasn't really random but followed some pattern, in which case it isn't fundamentally different from a verbose cipher.
Compression algorithms (like the well-known 'zip') specifically increase the entropy, and the uncompression decreases it.
To decode the Voynich MS text, one is then looking for something similar to compression.
Removal of nulls would be a form of compression, and combining characters another.
Things I have seen in some translation attempts, e.g. suggestions that the gallows characters represent the Nahuatl character pair 'tl', are going into the wrong direction.....
Another suggestion: that the text is vowelless, and vowels need to be inserted, is equally going in the wrong direction. Such a further reduction of entropy would need to be balanced by 'something else' that stronly increases it.
Surely the answer, or at least part of it, is that we're deconstructing characters, or character combinations, too much?
I know that the issue over [ckh, cth, cfh, cph] has been mentioned before, but also [ch, sh], [e, ee, eee], and [in, iin, iiin]. What's the highest entropy we can reach just by making different (but reasonable) character identifications?
Also, what improvements would it make if we assumed, for the sake of argument, that some characters had more than one permissible reading? The readings could be linked (such as voiced and voiceless) and distinguished in speech though not in writing.
RE: The Voynich Manuscript is a cipher because...? - ReneZ - 25-02-2016
The most anomalous, and therefore indicative value is that of h2, which tells about the 'unpredictability' of how characters combine, though it has to be seen in relation with h1, which is the unpredictability of single characters.
The Currier alphabet is the most 'compressed' we have, and gives the highest entropy. Taking some values from this page:
You are not allowed to view links. Register or Login to view.
Currier and Eva transcriptions of herbal A text have almost identical h1 values of 3.8, though Currier has 36 symbols and Eva only 26.
The h2 in Eva is 1.99 and in Currier 2.32.
For a Latin text, h1 and h2 values are 4.0 and 3.3 respectively.
This clearly shows the magnitude of our problem.
By the way, and importantly: a compression algorithm increases the entropy because during the compression there is no loss of information.
The entropy per character increases because the same amount of information is distributed over fewer characters.
Whatever the Voynich author did that ultimately reduced the entropy may have been a spreading of information over more characters, but it could also have been a reduction (loss) of information.