In Koen's most recent video, he interviewed an expert in historical constructed languages. She mentioned the Lingua Ignota, Enochian, and even more recent examples up to the 18th century. Her argument was that generally those examples feature statistical patterns that are more consistent with natural languages, for example Lingua Ignota mapping directly in syntax to Latin.
But what I found lacking was the discussion of "mathematical", "philosophical" and a-priori constructed languages that gained huge popularity during the 1600s. Wilkins, Leibniz and several others all made or proposed highly structured mathematical conlangs which group all things into a hierarchy and have very well defined, and often alien grammar and semantic systems.
What is to say that the constructed language of the Voynich can't be a constructed language much more similar to these examples than Lingua Ignota?
It would be anomalous for a pre-1600s thinker to devise such a strange and structured mathematical conlang, but everything about the Voynich is already anomalous. It's not a stretch to believe the author likely had a very different thought process to most people of their era, if not most people in all of history.
I think this idea should be taken seriously as an avenue of research, and compare the Voynich Manuscript more to languages like Wilkin's ordered mathematical language, or even modern examples line Ithkuil, than historical examples of constructed languages that closely resemble natural languages, of which there are already very few to compare
It's surely a possibility, and a constructed/philosophical language may possibly explain some statistics of the text, ie., the word structure. The problem is cracking it in a clear, demonstrable way (it's unfortunately rather easy to assign arbitrary meanings to glyphs groups and reconstruct a more or less coherent, and fully invented, text/meaning).
(10-10-2025, 08:44 AM)Mauro Wrote: You are not allowed to view links. Register or Login to view.It's surely a possibility, and a constructed/philosophical language may possibly explain some statistics of the text, ie., the word structure. The problem is cracking it in a clear, demonstrable way (it's unfortunately rather easy to assign arbitrary meanings to glyphs groups and reconstruct a more or less coherent, and fully invented, text/meaning).
I think the only way to truly verify any system, as many Voynich researchers have pointed out, is not simply to have the "translation" of pages, using arbitrary meaning assignments, but instead to identify the text generation method.
If someone hypothetically invented their own constructed language of a very similar type, using the same style or mathematical mode of thinking as the original author, their own constructed language may have many similarities to Voynichese in terms of its fundamental rules and structure. If this language happened to be extremely similar to Voynichese, but functioned with the same strange paradigm of mathematical constructed language, they would be able to write coherent sentences that would potentially have similar statistical properties.
What if this hypothetical modern constructed language was similar enough to Voynichese that it elucidated the exact logic behind the text generation method, as well as the possible thinking of the motivation and philosophical underpinnings of why the language is generated in such a way?
While this could still be considered arbitrary, if their system was able to predict specific ways in which the text is generated, for example by predicting the exact order of pen strokes, make accurate predictions about word and line rules, and predict which one-off rare glyphs would be likely to occur, that seems like it would be significant evidence that this system would be uncovering something more fundamental than typical arbitrary meaning assignments. Especially if this system could also produce new Voynichese text which matches all statistical patterns, but also encodes genuinely coherent meanings in a consistent system.
I think if this system existed, and worked as described, it would be quite a validation of the Constructed Language hypothesis.
The reality is that we still don't know what Voynichese is. Simple cipher, complex cipher, glossolalia, generated nonsense, code or a constructed language with mathematical, mystic, religious or philosophical underpinning. We just don't know.
I've noticed when I talk about hypothetical scenarios with scholars, they always think in terms of likelihood. So when Claire says it's probably not a constructed language, this means that the likelihood of it being a constructed language is very low given our knowledge of the MS and the nature of constructed languages around the 15th century.
The fact that she involves later examples strengthens the argument in a way, because it shows that even after the VM was made, the tendencies in constructed languages were still different.
Now again, we don't know what it is, and there is clearly nothing like it. So it could be something more mathematical, but she doesn't think of this as very likely.
To be honest, I don't think anyone knows what is most likely based on historical precedent though.
I very much like the idea of a philosophical conlang, and don’t think such a concept is anachronistic at all. How familiar are you with Rev. Ramón Llull and his Ars series (1274~1308)? This body of work is the next big [possibly] Voynich-related rabbit hole I plan to dive down. Llull claimed that the algorithm he invented, expressed largely in flowcharts and volvelles (!), could be used to guide someone toward the answer to really any question about life and the larger world. That’s a pretty eyebrow-raising claim — any work of art or scholarship that claims to be about everything is typically not about much of anything. I’m intrigued though, and willing to give it a chance.
Father Llull’s main use of his invention was convincing Muslims that they should convert to Christianity. He almost lost his life to an angry mob, trying this at a piazza in Tunis. So I can see why Father Llull is regarded by some today as a bit … touched, and his magnum opus of questionable philosophical merit. Llull’s Ars series is nevertheless very valuable historically, to people interested in the history of divination techniques and their cultural diffusion from the Arab world. Just like alchemy yielded what we know today as chemistry, I imagine a historical line can be drawn between popular book- and volvelle-based divination systems, and the design of computational and logical problem solving algorithms as we know them today.
Father Llull died in 1316. By that time he had traveled extensively, and founded academic departments (mostly for Semitic languages — he was a very talented linguist) at a number of renowned universities across Europe. He was well-heeled and well connected, and doubtless had a deep impact on a number of contemporaries. It’s not at all inconceivable to me that a student or protégée of Llull’s could have thought to continue his legacy by condensing his Ars into a philosophical conlang, with each glyph symbolically representing a specific movement or state of the algorithm, and each word's (or each line’s) sequence of glyphs representing a specific path through the algorithm. If Llull’s Ars really could be used like a library book classification system to describe, classify, and map all important aspects of the human experience, I can’t see why it couldn’t form the basis of an a priori constructed philosophical language or notation system. Probably a surprisingly productive and versatile one too, in terms of the range of meanings expressable and encodable. That said, such a conlang would likely suffer the same problem as most a prior philosophical conlangs: steep learning curves, unwieldy phonology, and unnatural syntax, and therefore kind of too impractical for most people to use. (Anyone here ever try to learn Lojban?)
As I make my way through Llull’s Ars, I’m going to keep in mind all that I’ve read and discussed about the properties and structure of Voynichese, and see if something clicks in my mind. I’ll probably also be rereading Patrick Feaster’s experiments with designing a cipher that fits Voynichese as we have it, and David Jackson and Brian Cham’s exploration of volvelles.
(10-10-2025, 01:58 PM)Koen G Wrote: You are not allowed to view links. Register or Login to view.The reality is that we still don't know what Voynichese is. Simple cipher, complex cipher, glossolalia, generated nonsense, code or a constructed language with mathematical, mystic, religious or philosophical underpinning. We just don't know.
I've noticed when I talk about hypothetical scenarios with scholars, they always think in terms of likelihood. So when Claire says it's probably not a constructed language, this means that the likelihood of it being a constructed language is very low given our knowledge of the MS and the nature of constructed languages around the 15th century.
The fact that she involves later examples strengthens the argument in a way, because it shows that even after the VM was made, the tendencies in constructed languages were still different.
Now again, we don't know what it is, and there is clearly nothing like it. So it could be something more mathematical, but she doesn't think of this as very likely.
To be honest, I don't think anyone knows what is most likely based on historical precedent though.
I think in this matter in particular, expertise in historical constructed languages tells us essentially nothing about the Voynich in particular. As you said, we don't know. And we cannot assess likelihoods of something we fundamentally don't understand.
However, I would like to highlight that in your video, your expert did go over some examples of historical conlangs, including more recent examples, what I thought was notably missing, and an important piece of the puzzle, were the conlang renaissance in the 1600s in England.
We have at least 5 major examples of different ideologies for mathematical, philosophical constructed languages of that time period.
Francis Lodwick - 1652, The Groundwork or Foundation laid (or So Intended) for the Framing of a New Perfect Language
Thomas Urquhart - 1653, Logopandecteision
George Dalgarno - 1661, Ars signorum
John Wilkins - 1668, An Essay towards a Real Character, and a Philosophical Language
Gottfried Leibniz - 1678, Lingua generalis
Each of these have different philosophies and ideals for how to construct a philosophical, universal language from first principles reasoning.
Although we don't know for certain, at all what the Voynich Manuscript is, and whether it is one of these languages, like all scientific endeavors, it's a hypothesis that should be tested. With the Voynich, we have at least ruled out simple substitution ciphers of Latin, a hoax by Voynich himself, or pure glossolalia with no purpose or structure whatsoever.
We should take seriously the idea that it is likely enough to be a philosophical a priori language that we should do serious investigation testing whether this could be true, and seeing if there are parallels between any historical or modern examples of such languages.
If we're deferring purely to expertise, William Friedman, who studied the Voynich Manuscript for a decade said
"The Voynich Manuscript was an early attempt to construct an artificial or universal language of the A-Priori type" —Friedman
I think it's reasonable to consider it a serious possibility and a line of inquiry worth exploring further.
(10-10-2025, 03:27 PM)RenegadeHealer Wrote: You are not allowed to view links. Register or Login to view.I very much like the idea of a philosophical conlang, and don’t think such a concept is anachronistic at all. How familiar are you with Rev. Ramón Llull and his Ars series (1274~1308)? This body of work is the next big [possibly] Voynich-related rabbit hole I plan to dive down. Llull claimed that the algorithm he invented, expressed largely in flowcharts and volvelles (!), could be used to guide someone toward the answer to really any question about life and the larger world. That’s a pretty eyebrow-raising claim — any work of art or scholarship that claims to be about everything is typically not about much of anything. I’m intrigued though, and willing to give it a chance.
Father Llull’s main use of his invention was convincing Muslims that they should convert to Christianity. He almost lost his life to an angry mob, trying this at a piazza in Tunis. So I can see why Father Llull is regarded by some today as a bit … touched, and his magnum opus of questionable philosophical merit. Llull’s Ars series is nevertheless very valuable historically, to people interested in the history of divination techniques and their cultural diffusion from the Arab world. Just like alchemy yielded what we know today as chemistry, I imagine a historical line can be drawn between popular book- and volvelle-based divination systems, and the design of computational and logical problem solving algorithms as we know them today.
Father Llull died in 1316. By that time he had traveled extensively, and founded academic departments (mostly for Semitic languages — he was a very talented linguist) at a number of renowned universities across Europe. He was well-heeled and well connected, and doubtless had a deep impact on a number of contemporaries. It’s not at all inconceivable to me that a student or protégée of Llull’s could have thought to continue his legacy by condensing his Ars into a philosophical conlang, with each glyph symbolically representing a specific movement or state of the algorithm, and each word's (or each line’s) sequence of glyphs representing a specific path through the algorithm. If Llull’s Ars really could be used like a library book classification system to describe, classify, and map all important aspects of the human experience, I can’t see why it couldn’t form the basis of an a priori constructed philosophical language or notation system. Probably a surprisingly productive and versatile one too, in terms of the range of meanings expressable and encodable. That said, such a conlang would likely suffer the same problem as most a prior philosophical conlangs: steep learning curves, unwieldy phonology, and unnatural syntax, and therefore kind of too impractical for most people to use. (Anyone here ever try to learn Lojban?)
As I make my way through Llull’s Ars, I’m going to keep in mind all that I’ve read and discussed about the properties and structure of Voynichese, and see if something clicks in my mind. I’ll probably also be rereading Patrick Feaster’s experiments with designing a cipher that fits Voynichese as we have it, and David Jackson and Brian Cham’s exploration of volvelles.
That is quite interesting. I haven't researched this conlang so that will be interesting to learn about.
I think your point about the steep learning curves of traditional conlangs is really interesting, and this is something I've thought about in relation to the VMS.
In many examples of conlang construction, even the originators of the languages themselves hardly are fluent in the language, for the reasons you highlighted. Clunky, unintuitive rules, unwieldy phonology, steep learning curves, etc.
If we are to take seriously the hypothesis that the VMS is a writing in a conlang, the sheer fluidity, depth and breadth of writing in one's own conlang is already anomalous. It could tell us something about the nature of the VMS conlang, that if it is a conlang, that conlang would have to be easy enough to learn and speak, and really think in, that someone could write hundreds of pages in it, or even so easy to learn that an entire group of people learned it if we are to take the idea seriously of multiple scribes.
In that case, it may rule out certain classes of philosophical conlangs which have exceedingly hard to memorize and complex rules. For example, the VMS likely doesn't have the same philosophical structure as say, Ithkuil, which is a conlang that's notoriously nearly impossible for anyone to learn to use practically.
I think this supports the idea that if the VMS is a conlang, it's likely remarkably simple to learn, and fairly straightforward in its structure, as unintuitive as that may seem on the surface.
Lately I’ve been trying to look at the Voynich manuscript in a more objective way, not to solve it, but to understand how its text behaves. Along the way I’ve been reading about two Catalan figures from the 13th and 14th centuries, Ramon Llull (like RenegadeHealer) and Arnau de Vilanova, and I keep thinking there might be a link between their worlds.
Llull built a logical system made of letters and rotating figures that could combine ideas and represent universal truths. Arnau was very different, more of a doctor and alchemist, writing about herbs, baths, astrology and how nature affects the body. There are some books attributed to him but not confirmed, but at least there was a kind of follow up school of.his thoughts, keeping his way of thinking alive. His genuine works were usually not much illustrated (at least what I have seen in internet). That made me think: what if the Voynich were something like an “illustrated Arnau” - the same kind of medical and natural knowledge, but shown visually to transmit it more clearly, as Lisa Fagin Davis suggested in her recent talk? (Meaning sort of school papers to learn)
If someone took Arnau’s themes and expressed them through Llull’s kind of symbolic, rule-based structure, the result could look surprisingly close to the Voynich. The herbal drawings, the bathing scenes, the cosmic diagrams, and the systematic, almost algorithmic writing would all make sense in that mix.
Of course, it’s only a personal theory from an amateur, just an idea that helps me explore the manuscript with more curiosity than certainty.
qoltedy Wrote:I think this supports the idea that if the VMS is a conlang, it's likely remarkably simple to learn, and fairly straightforward in its structure, as unintuitive as that may seem on the surface.
I agree, and have had this fleeting thought before: that the VMS’s encoding of information is actually a lot simpler than many of us who’ve puzzled over minute details of the text have ever considered. This would obviate the endless discussions over how many glyphs there are, what to do about ambiguous glyphs, whether the shape and placement of the top curve on EVA=[sh] or the curvedness or straightness of the tail on EVA=[p], are meaningfully significant, and other little details we like to sweat, hoping to find some clue. Because if the VMS’s text is indeed a simple (but not intuitive) conlang that’s easy to learn (but not intuitive) by design I can’t imagine these minor details matter or carry any information. Such a scenario favors a writing system that can be learned easily, and is fairly forgiving of errors and ambiguities, when it comes to comprehension. A modern, popular, and highly user-friendly minimalist conlang like Toki Pona comes immediately to mind.
This whole “Oh wow, that was it?! It was just that simple all along?!” kind of solution scenario also applies to @Jorge_Stolfi’s theory that the VMS text is an idiosyncratic conscript for a natural language, likely a wholly unwritten one, dictated by a native speaker to the scribes. I’ve studied Mandarin Chinese, and although I don’t think there’s a high chance the underlying dictated language was Sinitic, I have to say, the Voynichese writing system could easily be repurposed to phonetically transcribe Mandarin. Chinese is a highly isolating / analytical language, with most words one or two syllables, each syllable mappable to a Chinese character. Every Chinese syllable has four important properties:
- A rime, consisting of a simple vowel, diphthong, or triphthong, from a limited list
- One of four tones: high, rising, low, or falling
- An onset consonant, which can be absent, and when present, constrains which rimes can follow (but not which tone or coda)
- A coda, limited to -n or -ng, absent about half of the time, and not permissable after a triphthong rime.
I could see how each of these aspects of a Chinese syllable could be encoded by one of the rigid properties of a vord. For example, could an initial EVA=[o] or [qo], absent about half the time, be the way to encode the coda -n or -ng? Could the combinations of upward and downward curves and horizontal bars be the way to encode the tone? The series of EVA=[e]s or EVA=[i]s ways of representing the vowel combinations of the rime? I’m making this up. But you get the idea. If the scribes were taking dictation from illiterate speakers of a fairly analytical and highly exotic language, whose phonology assigned four important and noticeable aspects to every syllable that needed to be encoded, one might come up with a way to faithfully represent it in writing in a way that looked a lot like Voynichese.
quimqu Wrote:Arnau de Vilanova
I’m not familiar with him at all. Thank you for the lead, I’ll definitely be doing some reading about this chap.
(10-10-2025, 04:54 PM)quimqu Wrote: You are not allowed to view links. Register or Login to view.Lately I’ve been trying to look at the Voynich manuscript in a more objective way, not to solve it, but to understand how its text behaves. Along the way I’ve been reading about two Catalan figures from the 13th and 14th centuries, Ramon Llull (like RenegadeHealer) and Arnau de Vilanova, and I keep thinking there might be a link between their worlds.
Llull built a logical system made of letters and rotating figures that could combine ideas and represent universal truths. Arnau was very different, more of a doctor and alchemist, writing about herbs, baths, astrology and how nature affects the body. There are some books attributed to him but not confirmed, but at least there was a kind of follow up school of.his thoughts, keeping his way of thinking alive. His genuine works were usually not much illustrated (at least what I have seen in internet). That made me think: what if the Voynich were something like an “illustrated Arnau” - the same kind of medical and natural knowledge, but shown visually to transmit it more clearly, as Lisa Fagin Davis suggested in her recent talk? (Meaning sort of school papers to learn)
If someone took Arnau’s themes and expressed them through Llull’s kind of symbolic, rule-based structure, the result could look surprisingly close to the Voynich. The herbal drawings, the bathing scenes, the cosmic diagrams, and the systematic, almost algorithmic writing would all make sense in that mix.
Of course, it’s only a personal theory from an amateur, just an idea that helps me explore the manuscript with more curiosity than certainty.
I think this is quite a compelling way to look at it, and I've come to relatively similar conclusions through a different logical path.
Assuming the conlang system is relatively simple, and the perceived complexity emerges from combinations of simpler primitives, I would expect whatever those morphemes are, which combine to make new meaning, would be very universal ideas. Perhaps so universal, that there is a remarkably small number of them, encoding a remarkably small amount of fundamental concepts.
I would suspect that a mind capable of creating such an elegant but functional system would likely be what we today call a "genius". Likely someone with a scientific, mathematical intuition and very much creativity and originality, as evidenced by well, everything in the Manuscript.
Such a Medieval genius would very possibly be interdisciplinary/polymathic, and be interested in a number of scientific or technical questions. The diagrams absolutely have an explanatory flavor, with the labels, seemingly describing some real or imagined process in some system we have yet to understand. I personally think that such a genius would use visuals to aid their explanations and proto-scientific musings. Perhaps this conlang was not just the language that they wrote in, but the language that they THOUGHT in. What if it helped them to more clearly understand or visualize certain theories by using their constructed language rather than the natural language(s) they almost certainly knew?