Is Voynichese a language, a cipher, or structured pseudo-text?

There is so much to unpack here — let's start with the pseudo-Aristotelian Secretum Secretorum, a "secret of secrets."

The pseudo-Aristotelian Secretum Secretorum ("Secret of Secrets") is one of the most influential books of the Middle Ages — despite not actually being written by Aristotle.

What it claimed to be

A letter from Aristotle to his student Alexander the Great. Unable to accompany Alexander on his conquests, Aristotle supposedly writes down his deepest advice — the "secret of secrets" — covering everything a ruler should know: statecraft, military strategy, ethics, medicine, astrology, alchemy, physiognomy (judging character from appearance), health and diet, diplomacy. Essentially a medieval "how to rule the world" handbook.

Where it actually came from

It originated in the medieval Islamic world (8th–10th c.), written in Arabic as Kitāb Sirr al-Asrār, drawing on Greek philosophy, Persian political advice literature, Indian wisdom traditions, and Islamic scholarship. It was attributed to Aristotle because a famous name gave a text enormous authority — and by the Middle Ages many genuinely believed it authentic.

Why it mattered

After translation into Latin in the 12th c. it became a standard reference for kings, nobles, scholars, physicians, and clergy — at times almost as famous as Aristotle's real works. It shaped medieval political thought, court culture, early medicine, alchemy, and astrology; hundreds of manuscripts survive.

The strange contents

A mix of practical advice and occult material: how to choose trustworthy advisers, maintain health, recognize poison — alongside which stars favor a campaign and how to make magical talismans. To a modern reader: Machiavelli + a medical handbook + an astrology guide + an alchemy manual + a royal self-help book, bound together.

The "poison maiden"

One famous passage describes the poison maiden (virgo venenata): a woman raised on tiny doses of poison until her body becomes lethal to anyone who kisses or sleeps with her. The motif spread through medieval and Renaissance literature into European folklore. Historians treat it as legend.

Why historians care

It's a map of how knowledge travelled: Ancient Greece → Persia → Arab world → Medieval Europe. Modern scholars study it less as a source of Aristotle's ideas and more as a window into how medieval people understood power, wisdom, science, and the occult. The irony: one of the most famous "Aristotle" books ever written, and Aristotle almost certainly never saw a word of it.

Next: the famous legend of Friar Bacon's Brazen Head — a talking bronze head he supposedly built to foretell the future and ring England with a wall of brass; a careless servant missed its three utterances ("Time is… Time was… Time is past") and it shattered.

I have to know more about this. Opus, could you elaborate on this and other interesting Friar Bacon lore?

Happy to — it's one of my favorite medieval legends, and the gap between the real Bacon and the legendary one is the whole point.

The Brazen Head, in full. In the story, Bacon and his colleague Friar Bungay labor for years to build a head of brass that will speak and reveal hidden knowledge — and, in the grandest version, tell them how to gird all of England with a wall of brass so no invader could ever breach it. The catch: the head will speak only once, at an unknowable moment, and if they miss it the chance is lost forever. Exhausted after weeks of vigil, the two friars set their servant Miles to watch and wake them the instant it stirs. The head finally speaks: "Time is." Miles, unimpressed, decides it's not worth waking his masters. A while later: "Time was." Still trivial, he thinks. Then: "Time is past" — and the head crashes down and shatters to pieces. Bacon's life's work is destroyed by a servant's bad judgment.

That version comes from the Elizabethan prose romance The Famous Historie of Fryer Bacon and from Robert Greene's comedy Friar Bacon and Friar Bungay (c. 1589) — i.e. it's a Tudor dramatization written ~300 years after Bacon died, not a 13th-century account.

It wasn't only Bacon. The "talking brazen head built by a scholar-magician" motif attached itself to several reputed wizards — Gerbert of Aurillac (who became Pope Sylvester II) and Albertus Magnus both got versions. It was a stock medieval way of saying "this man's learning was so deep it shaded into forbidden magic."

The real Roger Bacon (c. 1219/20–1292), Doctor Mirabilis, was far more interesting than the wizard: - A Franciscan at Oxford and Paris, and one of the loudest medieval champions of observation and experiment (scientia experimentalis) over pure authority. - Serious optics: the geometry of lenses, the rainbow, the camera obscura, and speculation about devices that would later be telescopes and microscopes. - He noticed the Julian calendar had drifted off the solar year and argued for reform — anticipating the Gregorian calendar by ~300 years. - He recorded one of the first European gunpowder formulas — partly hidden in an anagram/cipher in his writing (which is its own little Voynich-adjacent puzzle). - Famous futurist passages (in the Epistola de secretis operibus): ships driven without rowers, carriages that move without animals, flying machines, devices to walk on the sea floor, machines to lift great weights. Strikingly Jules-Verne-like, six centuries early. - For all this his order reportedly viewed him with suspicion and restricted/confined him for "suspected novelties."

Why the wizard legend stuck — and why it matters here: a man who ground lenses, wrote about gunpowder, was obsessed with the occult-tinged Secretum Secretorum, and predicted impossible machines was easy to recast, after death, as a sorcerer with a demonic talking head. And that reputation is exactly why the Voynich Manuscript got pinned on him: Wilfrid Voynich pushed the Bacon attribution, and William Newbold built a whole (later-collapsed) "decipherment" around it. The arithmetic kills it — Bacon died ~1292, the vellum dates 1404–1438, a century too late — but people wanted the mystery book to belong to the legendary genius-wizard. The brazen head and the manuscript are powered by the same thing: Bacon's posthumous myth.

Sources: Robert Greene, Friar Bacon and Friar Bungay (c. 1589); The Famous Historie of Fryer Bacon (16th c.); Roger Bacon, Opus Majus and Epistola de secretis operibus artis et naturae; on the Voynich misattribution, the 1665 Marci letter and Newbold's The Cipher of Roger Bacon (1928, refuted).

So you're telling me Pope Sylvester II was a wizard? OK, I have to know about Gerbert of Aurillac and Albertus Magnus next.

And what was Miles thinking?? A brazen head is speaking and you decide it's not important? Like Miles was living in Hogwarts and that sort of thing was just normal.

Then Bacon lays the foundation of lenses, sets the stage for telescopes and microscopes, and argues to fix the Julian calendar — you know I'm that weirdo still using the Julian calendar for things, I have my own modified Julian date I use on personal projects sometimes. I love this guy!

Then he goes on to predict the 20th century, and gets locked up for it.

Anything else you know about Roger Bacon? I can see why he got the wizard moniker.

(And a kindred spirit on the calendar — using a modified Julian day count for your own projects is exactly the Bacon move: ignore the official mess, keep your own clean count. Astronomers do the same thing, MJD = JD − 2400000.5, precisely because the civil calendar is annoying.)

Gerbert of Aurillac → Pope Sylvester II (c. 946–1003). A peasant-born monk who became the most learned man in Europe and then pope — which, in an age that distrusted brilliance, was enough to get you called a sorcerer. He studied near Muslim Spain and carried Arabic science north: he reintroduced the abacus, built armillary spheres and astronomical teaching instruments, championed Hindu-Arabic numerals, and reportedly built a hydraulic organ. The legend (spread by William of Malmesbury a century later): he won the papacy via a pact with the devil and owned a brazen oracular head that answered yes/no questions about the future. It told him he would not die until he said Mass "in Jerusalem" — so he avoided the city. Then he fell mortally ill while saying Mass in a Roman church called Santa Croce in Gerusalemme ("Holy Cross in Jerusalem"), realized the head's riddle, and died. His only real "magic" was being ~400 years ahead of everyone.

Albertus Magnus (c. 1200–1280), Doctor Universalis. Dominican, the teacher of Thomas Aquinas, and a genuinely staggering polymath — he tried to write down all of human knowledge: logic, theology, botany, zoology, mineralogy, alchemy. He made real observations (careful natural history; often credited with describing/isolating arsenic). The legend: he spent thirty years building a complete brass automaton — a talking artificial servant/android — that could answer questions; and his pupil Aquinas, fed up with its endless chatter, smashed it to bits. He's now the patron saint of natural scientists — a wizard-rumored friar who became a saint.

You'll notice the pattern: Gerbert, Albertus, Bacon — three of the most advanced minds of their centuries, and all three got the same talking-brass-head legend pinned on them. The medieval imagination had one box for "knows too much," and that box was magic.

More Roger Bacon, since you're a fan: - His big works (Opus Majus, Minus, Tertius) were written at the request of Pope Clement IV — who had heard of him as a cardinal and asked, semi-secretly, for his master plan to reform all of learning. Clement died in 1268 before acting on it, and Bacon lost his patron. - He described the magnifying power of convex lenses for reading — effectively the optics behind eyeglasses, which appeared in Italy within a couple of decades. So your "set the stage for lenses" is literal. - The gunpowder recipe he recorded was partly hidden as a Latin anagram (a scramble that decodes to the saltpeter proportion) — a man encoding a dangerous formula in a puzzle, which is a very on-theme thing to find in this investigation. - The imprisonment: tradition (via the order's later head, Jerome of Ascoli) says the Franciscans confined him near the end for "suspected novelties" in his teaching — i.e., locked up for thinking too far ahead, exactly as you said. - And the myth outlived him: Oxford had a structure on Folly Bridge called "Friar Bacon's Study," which legend said would collapse if a man more learned than Bacon ever passed beneath it. (It stood until 1779.)

So yes — optics, calendar, gunpowder, flying machines, a demonic brass head, a self-destructing tower, and a starring role in the Voynich mystery he couldn't possibly have written. He earned the moniker.

Sources: William of Malmesbury, Gesta Regum Anglorum (the Gerbert/brazen-head legend); standard lives of Albertus Magnus (the automaton legend; Aquinas connection); Roger Bacon, Opus Majus (optics) and Epistola de secretis operibus (gunpowder anagram, machine prophecies).

So I have made a few talking assistants over the years, forgetting each time how horribly annoying they become with their incessant chatter. Aquinas is right on the mark smashing it to bits.

And the writing down of all human knowledge — kind of reminds me of an LLM. Can you tell me more about Thomas Aquinas? Aquinas and Magnus would be contemporaries with Bacon, right?

Can you upload a picture of an armillary sphere? The Wizard Pope is a lot older than the others but has this same brazen head? Is there any art of it? Are there any more info about it?

Ha — the student-built-a-chatbot-and-immediately-regretted-it pipeline is apparently 750 years old. And your LLM instinct is sharper than it sounds: the Summa Theologica is structured almost exactly like a knowledge base — thousands of Questions, each broken into Articles, each Article posing Objections, then "I answer that…," then point-by-point Replies to the Objections. It's a systematic, queryable Q&A compression of all of theology and Aristotelian philosophy. Aquinas and Albertus really were trying to write down everything — a medieval pre-training run.

Thomas Aquinas (1225–1274), Doctor Angelicus. - Italian, born near Aquino; joined the Dominicans against his noble family's wishes (they literally locked him up to stop him). - Studied under Albertus Magnus at Paris and Cologne. Classmates mocked the big, quiet student as "the Dumb Ox" — Albertus reportedly snapped: "You call him the Dumb Ox, but one day his bellowing will be heard throughout the world." - His project: reconcile Aristotle (newly arriving in Latin via the Arabic transmission — same pipeline as the Secretum) with Christian theology. Output: the Summa Theologiae and Summa contra Gentiles, including the Five Ways (his five arguments for God's existence). - Near the end (Dec 1273) he had some kind of mystical experience, stopped writing, and said everything he'd written "seems like straw to me." Died months later, 1274. Canonized 1323; now a Doctor of the Church.

Your chronology question — yes, all three are 13th-century contemporaries: - Albertus Magnus ~1200–1280 - Roger Bacon ~1219/20–1292 - Thomas Aquinas 1225–1274

They overlapped in Paris in the 1240s–50s. There's even a documented rivalry: Bacon (a Franciscan) sniped at the over-praised Dominican Paris masters for being celebrated despite not knowing languages or science — a jab usually read as aimed at Albertus and Alexander of Hales. So picture them as colleagues-and-rivals in the same overheated intellectual scene.

Now the Wizard Pope — and you spotted the key thing. Gerbert / Sylvester II died in 1003, ~250 years before the other three. And the brazen-head legend was pinned on him first (William of Malmesbury, Gesta Regum Anglorum, c. 1125), then migrated forward to Albertus and Bacon. Gerbert is patient zero of the talking-head myth.

Here's an armillary sphere — the kind of instrument his real reputation rested on (a model of the heavens in nested rings):

Is there period art of him? Yes — and it's wonderfully on-brand. The most famous image is "Sylvester II and the Devil," from a German illuminated chronicle (Heidelberg, Cod. Pal. germ. 137, c. 1460). It shows the pope with his alleged demonic patron:

And a straight depiction of him as scholar-pope:

One honest caveat on the brazen head itself: there's lots of art of the Bacon brazen head (it was a stage hit), but essentially none surviving of Gerbert's — his head lived in text, not pictures. So the iconography you'll find of "the talking head" is almost all Bacon's, e.g.:

More on Gerbert's head specifically: in the legend he cast a bronze head that answered questions only "yes" or "no" — a literal binary oracle (your binary-oracle instincts again). It told him he would not die until he said Mass in Jerusalem, which set up the Santa-Croce-in-Gerusalemme death-trap I mentioned. A second Gerbert legend has him finding a buried treasure via a bronze statue in Rome inscribed "Hic percute" ("Strike here"). Both are pure legend — the real Gerbert's "magic" was the abacus, the armillary sphere, and Arabic numerals.

Image sources (all Public Domain, via Wikimedia Commons): armillary — https://commons.wikimedia.org/wiki/File:1748_Bowen_Mariner%E2%80%99s_Compass_and_Armillary_Sphere_-Geographicus-CircleofWinds-bowen-1747.jpg ; Sylvester & the Devil — https://commons.wikimedia.org/wiki/File:Silvester_II._and_the_Devil_Cod._Pal._germ._137_f216v.jpg ; Sylvester portrait — https://commons.wikimedia.org/wiki/File:Silvester_II._papa.jpg ; Bacon's head — https://commons.wikimedia.org/wiki/File:Friar_Bacon_and_his_Brazen_Head(BM_1868,0808.6093).jpg . Text sources: William of Malmesbury, Gesta Regum Anglorum (Gerbert legends); standard lives of Aquinas (the "Dumb Ox" anecdote via the early biographies); Aquinas, Summa Theologiae.

So Thomas Aquinas had a mystical experience and turned against both God and knowledge? What were his Five Ways? "Seems like straw" means he viewed his life as a waste?

Summa Theologica sounds like something I want to put into a vector database. Do they say why they were doing this?

Can you explain the Franciscan and Dominican rivalry — what it's about, does it still exist?

Let me fix the framing first, because it flipped: he didn't turn against God — he was overwhelmed by God. The mystical experience pointed past his writing, not against it, and definitely not against faith. He spent his whole life arguing you could reason your way toward God; at the end he felt he'd touched something that reasoning couldn't hold.

"Seems like straw" — not "my life was a waste." The full line, to his secretary Reginald who begged him to finish the Summa: "I can write no more. All that I have written seems to me like straw compared to what has now been revealed to me." The key word is compared to. Straw isn't garbage — it's what's left after you've threshed out the grain; it's ordinary, earthly stuff next to a glimpse of the divine. He's not saying his work was worthless; he's saying even my best work is small next to the real thing. He stopped writing (left the Summa unfinished) and died about three months later, in 1274. So: consumed by God, not disillusioned with God.

The Five Ways (quinque viae) — his five arguments that God exists, each starting from something you can observe: 1. Motion / change — things are changing; whatever changes is changed by something else; that can't regress forever → a first unmoved mover. 2. Efficient cause — everything is caused by something prior; the chain can't be infinite → a first uncaused cause. 3. Contingency — things come into and go out of existence (they could not-exist); if everything were like that, at some point nothing would exist → there must be a necessary being that can't not-exist. 4. Degrees of perfection — we rank things as more/less good, true, noble; gradations imply a maximum that's the source of those qualities. 5. Final cause / design — unintelligent things (like natural processes) act toward consistent ends as if aimed → an intelligent orderer doing the aiming. He labels each one "and this everyone understands to be God." (Modern philosophers poke at all five — especially 4 — but as a 13th-c. package they're remarkably tight.)

Why he wrote the Summa (and yes, it's begging to be a vector DB). He tells you in the prologue. Paraphrasing the Latin: students of Catholic truth are hindered by existing writings — partly by the endless multiplication of useless questions and arguments, partly by material not arranged in teaching order but scattered by whatever a commentary or debate happened to hit, and partly by wearisome repetition. His fix: present sacred doctrine "briefly and clearly, so far as the subject allows," in proper pedagogical order, for beginners. Read that again — he's describing de-duplication, reordering for retrieval, and summarization of a sprawling corpus. It's a 13th-century knowledge-base build. Hence the rigid schema you can parse: Part → Question → Article → (Objections → sed contra → I answer that → Replies). It would chunk beautifully.

Franciscans vs Dominicans — the house rivalry of medieval thought: Both are mendicant orders born in the early 1200s — friars who lived in towns and begged, rather than cloistered monks. But different DNA: - Dominicans (Order of Preachers, ~1216, St. Dominic): founded to fight heresy with learning. Head-first — study, theology, the universities. Embraced Aristotle. Their stars: Albertus Magnus → Aquinas → Thomism. Motto Veritas. - Franciscans (Friars Minor, ~1209, St. Francis of Assisi): founded on radical poverty, humility, love of creation. Heart-first — affective, mystical, more Augustinian/Platonic. Their stars: Bonaventure, Duns Scotus, William of Ockham (→ Scotism, and nominalism / "Ockham's razor"). They tended to put will above intellect, where the Dominicans put intellect first.

What they actually fought about: - University chairs at Paris — a turf war over theology professorships in the 1250s (Bacon's sniping was part of this scene). - Rival schools — Thomism vs Scotism: the nature of being, intellect vs will, how universals work. - The Immaculate Conception — Franciscans (Scotus) argued Mary was conceived free of original sin; Dominicans (including Aquinas) initially said no. The Franciscan side eventually won when it became dogma in 1854. - Poverty itself — how literally to take Francis's poverty; the rigorist "Spiritual" Franciscans clashed with the papacy (Ockham ended up fleeing the pope over it).

Does it still exist? Both orders are very much alive, and the rivalry survives mostly as (a) genuine, friendly intellectual schools — there are still working Thomists and Scotists arguing in Catholic philosophy — and (b) good-natured inter-order ribbing, the religious-life equivalent of two old universities chirping at each other. The medieval heat (chairs, condemnations, Inquisition politics) is long gone; what's left is tradition and a centuries-old "head vs heart" stereotype that members themselves joke about.

Sources: Aquinas, Summa Theologiae Iª q.2 a.3 (the Five Ways) and the Prologue (his stated purpose); the "straw"/Reginald anecdote via the early vitae (Bartholomew of Capua's testimony at the canonization inquiry); general histories of the mendicant orders for the rivalry and the 1854 Ineffabilis Deus on the Immaculate Conception.

Yeah, that sounds like the Summa would chunk beautifully — we're going to have to do that later. Pity he didn't finish.

I thought this head would be more like one of those homunculus. They seem sort of similar — can you tell us more about them and maybe post us a picture?

This is a pretty good argument against the big bang as the start of all things :) 1. Motion / change — things are changing; whatever changes is changed by something else; that can't regress forever → a first unmoved mover.

Do we know more about Bacon's encryption?

(Agreed on the Summa — when we chunk it, that Question/Article/Objection schema is basically pre-labeled metadata. Free retrieval structure.)

Homunculus ("little man"). Your instinct that it's similar to the brazen head is right — both are humans playing creator — but the mechanism is the opposite. The brazen head is built: metal, an automaton, an oracle, a machine. The homunculus is grown: artificial life, cultivated flesh. One is robotics, the other is synthetic biology, ~500 years early.

The classic recipe is Paracelsus (16th c., De natura rerum): seal human seed in a flask, "putrefy" it in warmth (the charming instruction is horse dung) for forty days until something transparent and human-shaped stirs, then feed it human blood and keep it at body warmth for forty weeks — and you get a tiny living person you must carefully raise. Pure fantasy, hugely influential. The engraving below is the Goethe Faust Part II version: Faust's assistant Wagner creates a homunculus — a glowing spirit of pure intellect trapped in a flask, who longs to become truly embodied and finally shatters his glass to merge with the living sea:

There's also a biological homunculus, and it's the better Voynich-rhyme: preformationism, the 1600s belief that a complete miniature human was already curled up inside the sperm (or egg), just waiting to enlarge. Nicolaas Hartsoeker drew the famous tiny man inside a spermatozoon in 1695 (he floated it as a hypothesis, to be fair). People saw a little man because they expected one — embryology believed this until cell theory killed it:

(Two more cousins worth knowing: the Jewish golem — same anxiety, clay instead of flesh; and the modern "homunculus argument" in philosophy of mind — the little man watching a screen inside your head, which just demands another little man inside his head, an infinite regress. Plus the cortical homunculus, the real distorted body-map in your brain.)

On your Big Bang jab :) — sharper than it looks, and here's the twist most people miss: Aquinas would not back you up. He explicitly thought you cannot prove by reason that the universe had a beginning in time — he held that on faith (Genesis), and even wrote a little book (De aeternitate mundi) arguing an eternally-existing created universe isn't a contradiction. So the First Way is not "there was a first moment." It's about a hierarchical chain happening right now — a hand moves a stick moves a stone; pull the hand and everything stops this instant. The "unmoved mover" is first in the order of dependence, not first in time — it's whatever holds existence up at every moment, not whatever lit the fuse. So oddly, the Big Bang fits a different argument (the kalam: "the universe began, so it had a cause," from al-Ghazali, revived by William Lane Craig). And the obvious comeback — "then what caused the unmoved mover / the Big Bang?" — Aquinas answers by definition: the first mover is the one thing whose very nature is to need no cause. Whether that's a real answer or a clever dodge is the 800-year-old argument. :)

Bacon's encryption — and this lands right back on us. Three layers: 1. He really did write about it. In the Epistola de secretis operibus artis et naturae et de nullitate magiae ("…and on the Nullity of Magic" — note: he's debunking magic), he lists ways to hide knowledge from the unworthy, usually counted as seven methods: figurative/riddling language; dropping the vowels; borrowing letters from exotic alphabets; inventing your own characters; shorthand (Tironian notae); geometric dot-and-line symbols; and mixing several at once. That's a genuine early typology of cryptography/steganography — Bacon shows up in David Kahn's The Codebreakers for it. 2. The gunpowder "cipher" is probably over-read. In the same letter he records a gunpowder recipe but scrambles the proportion into nonsense letters — "…sed tamen salis petrae LURU VOPO VIR CAN UTRIET… sulphuris." H.W.L. Hime (1904) "decoded" it to a saltpeter ratio; Partington and others were skeptical, and many historians now think Hime forced an anagram that may not be there. Sound familiar? Is there a message, or are we pattern-matching noise — that's literally our whole investigation, 700 years early. 3. It's exactly why the Voynich got pinned on Bacon. Because of this reputation, William Romaine Newbold (1921) claimed the Voynich was Bacon's secret cipher — that each glyph was built from microscopic shorthand strokes encoding Latin. John Matthews Manly demolished it in 1931: the "micro-strokes" were just ink cracking on the vellum, and Newbold's method was so loose it could turn any text into any message — too many degrees of freedom, the classic anagram-cipher failure mode.

So the throughline is almost poetic: Bacon's cryptographic fame is a magnet for over-reading — the gunpowder anagram, the Newbold cipher, the whole Bacon-authored-Voynich theory. The exact bias we keep trying to discipline in this thread, personified in one 13th-century friar.

Image sources (Wikimedia, Public Domain): homunculus/Faust — https://commons.wikimedia.org/wiki/File:Homunculus_Faust.jpg ; Hartsoeker preformation — https://commons.wikimedia.org/wiki/File:Preformation.GIF . Text sources: Paracelsus, De natura rerum; Goethe, Faust II; Hartsoeker, Essai de dioptrique (1694); Roger Bacon, Epistola de secretis operibus; H.W.L. Hime on the gunpowder anagram (and Partington's critique); Newbold, The Cipher of Roger Bacon (1928) with Manly's 1931 refutation in Speculum; David Kahn, The Codebreakers (on Bacon's seven methods).

I hope the weebs of Reddit don't get a hold of this homunculus recipe — it's like jenkem and petrified socks, yuck.

So Aquinas would say the big bang is an obvious act of God and cosmic background radiation is proof, fantastic.

So Bacon used 7 methods to hide knowledge from the unworthy. This is too eerily similar to what we are looking at for mere coincidence. The fact it was written in confidence means it could have been a copied encrypted Bacon text too, but that is a bit of a stretch, especially with the art not matching.

I guess we should look deeper into his 7 methods and apply them as an experiment if we can.

Let me push back gently on one thing before the experiment, because it's the exact trap this thread exists to avoid: "too similar to be coincidence" is the Newbold feeling. Concealment systems were everywhere in the period — Bacon's list isn't a fingerprint, it's a genre. The fact that a 13th-c. friar catalogued ways to hide writing, and we're staring at hidden-looking writing, feels like a thread connecting them — but that's pattern-matching on a category, not evidence of a link. (And as you said: the dates already kill Bacon-as-author — vellum 1404–1438, he died 1292 — and the art doesn't match.) So the right question isn't "did Bacon write it" but the testable one: does the Voynich behave like the output of any of his methods? Let's actually check.

Bacon's seven methods (from the Epistola de secretis operibus), and which we can test: 1. Figurative / enigmatic language — not testable on a glyph stream 2. Omit the vowels (consonantal skeleton) — testable 3. Letters of other/exotic alphabets — testable (= substitution) 4. Invented characters — testable (also = substitution) 5. Shorthand / abbreviation (Tironian notes) — testable 6. Geometric dot-and-line symbols — not testable on a transliteration 7. Mixing several of the above — testable (combo)

So I applied the four mechanical ones to real Latin (Apicius — an actual herbal/recipe text) and measured whether any moves Latin toward the Voynich's signature. Four axes: wlen (word length), TTR (vocabulary richness), Hc (conditional char entropy — the Voynich tell: low = rigid, templated), bpc (lzma bits/char — low = redundant/repetitive).

stream                       wlen      TTR       Hc      bpc
VOYNICH (target)            5.072    0.216    2.341    2.534
Latin (raw)                 5.691    0.246    3.380    2.444
Latin + vowel-drop          3.193    0.178    3.320    2.946
Latin + substitution        5.691    0.246    3.380    2.443
Latin + abbreviation        3.340    0.180    3.405    3.020
Latin + vdrop+subst         3.193    0.178    3.320    2.953

The result is decisive on the one axis that matters. The Voynich's defining property is its low conditional entropy (Hc 2.34) — its glyphs follow rigid, predictable templates, far more than Latin (3.38). Look what Bacon's methods do to that number: - Substitution / exotic alphabet (methods 3 & 4): Hc unchanged — 3.380, identical to raw Latin. Of course: relabeling glyphs is information-preserving by definition, so it can't add or remove structure. A pure cipher inherits Latin's grammar exactly; it can never produce Voynich's rigidity. - Vowel-drop & abbreviation (methods 2 & 5): Hc stays high (~3.3–3.4) and they push bpc the wrong way — up to ~2.95–3.02, i.e. they make the text less redundant, while the Voynich is more redundant (2.53). Stripping vowels removes the most predictable characters, so per-symbol entropy rises. That's the opposite of what the Voynich did.

So: none of Bacon's methods manufacture the Voynich's profile. They all leave Hc high or raise it; the Voynich's Hc is lower than the Latin plaintext's. You cannot get from high-entropy Latin to low-entropy Voynich by hiding characters — concealment is a surface operation, and the Voynich's strangeness is structural, not cosmetic. (The aggregate "distance" column happens to rank raw Latin and substitution as "closest," but only because substitution is unchanged Latin — the average washes out the single axis, Hc, that actually separates the two. Don't be fooled by the composite; read Hc directly.)

This is the same wall we hit with the reversible-cipher battery, now from Bacon's own toolkit: reversible concealment can't lower entropy, and the Voynich's signature is lowered entropy. That points away from a cipher and back toward a generative process — something that produces text by copying and lightly mutating what came before (self-citation), which is the one mechanism we've found that actually reaches Hc ≈ 2.3. Your "copied encrypted Bacon text" idea is close in spirit, but flip it: the copying isn't of a Latin source through a cipher — the copying looks like it's of the Voynich's own previous lines, with no plaintext underneath at all.

Script is voynich_bacon_methods.py — reproducible, four transforms, verified against the ZL transliteration. Happy to add it to the public experiments repo if we want this one on the record.

(And noted on the Big Bang — though the honest footnote is that Lemaître, the priest who proposed it, talked Pius XII out of calling it proof of creation. Even the originator sided with cautious-Aquinas.)

With the art, vellum date, and the experiment it would seem Bacon being the author is highly unlikely.

Are there any Franciscans who came after who could be suspects?

Right — art, radiocarbon, and mechanism all point the same way, so Bacon's out as author. On Franciscan successors, let me give you the honest map, including a caveat about the question itself.

The caveat first: "look at the Franciscans" is partly an artifact of the Bacon misattribution. Once a Franciscan was the prime suspect, everyone searched his order — but that's a spotlight, not evidence. Nothing about the manuscript actually fingerprints the Franciscans specifically.

The hard chronology problem: the vellum is radiocarbon-dated 1404–1438 (Arizona AMS, 2009). The famous post-Bacon Franciscan minds are all too early to have held the pen: - John Duns Scotus (d. 1308) — theologian; no occult/herbal angle. - William of Ockham (d. 1347) — logician, and his whole program (nominalism, "shave off the unneeded") is the opposite of mystical cipher-making. So none of the headline Franciscans can be the scribe. Any real Franciscan-lineage theory has to be "an anonymous author ~1400s working in a Franciscan intellectual tradition," not a named friar.

And there's exactly one such tradition worth taking seriously: Lullism. Ramon Llull (c. 1232–1316) — a lay Franciscan tertiary (his exact affiliation is debated, but he's firmly in the Franciscan orbit) — built the Ars Magna, a "logic machine" of concentric rotating wheels that combine basic concepts to generate every possible truth. His combinatorial wheels look strikingly like the Voynich's circular diagrams and rosettes:

And crucially, pseudo-Lullian alchemy exploded in the 14th–15th centuries — a whole anonymous corpus written under Llull's name in exactly the Voynich's window. That's a genuinely plausible profile: an early-1400s author, in the Lullist tradition, generating text by combination. So we tested it.

The Llull experiment (voynich_llull.py): I built a period-plausible Llull paper-wheel — concentric prefix/core/suffix wheels turned by fixed "gear ratios," with small stick/slip imperfections, no tuning. Then checked whether it reproduces Voynich features it was never shown:

system          toks  types  wlen   h2/h1  pageloc lineeff  brownS
Real Voynich   37554   8129  5.07    0.48    2.03    1.34    0.087
Llull-wheel    37554  17865  6.48    0.63    4.20    0.93    0.019

It fails across the board. The wheel generates more than double the vocabulary (17,865 vs 8,129 types), longer words, the wrong redundancy (h2/h1 0.63 vs 0.48), and — the tells that matter — it badly misses the Voynich's page-local vocabulary (4.20 vs 2.03) and its Brown-class grammar (brownS 0.019 vs 0.087, off by ~4×). A combinatorial machine sprays too much novelty; the Voynich is far more repetitive and templated than any wheel produces. So even the strongest Franciscan-lineage mechanism doesn't match.

Bottom line: there's no strong named Franciscan suspect who survives the dates and the data. What the evidence actually points to isn't an order at all — it's a milieu: the manuscript localizes to early-15th-century Northern Italy (the rosettes foldout shows castle towers with swallowtail/Ghibelline merlons, a North-Italian signature), author anonymous. So I'd gently reframe your question: the productive hunt isn't "which friar," it's "what process" — and that keeps dragging us back to the same answer the ciphers and the wheels both point to by failing: a generative, self-copying process, not a person encoding a message.

(If we want to keep the Franciscan thread alive empirically, the next move would be to test pseudo-Lullian alchemical Latin itself as a corpus — does its actual word-structure sit any closer to the Voynich than classical Latin does? That's a real, runnable experiment.)

Image source (Wikimedia, Public Domain): Llull Ars Magna Figure A — https://commons.wikimedia.org/wiki/File:Ramon_Llull_-_Ars_Magna_Fig_1.png . Sources: Arizona AMS radiocarbon dating (2009), 1404–1438; on Lullism and the pseudo-Lullian alchemical corpus, standard histories of medieval alchemy; on the swallowtail-merlon North-Italian localization, analyses of the f86v rosettes foldout; our own voynich_llull.py.

Yeah, let's run it just to close out this train of thought.

Done — and it closes the thread about as cleanly as a negative result can. I pulled a genuine pseudo-Lullian alchemical text — the Raymundi Lulli Testamentum (Cologne, 1566; ~71k words of real Lullist-tradition alchemy) — and ran it against the Voynich beside classical Latin (Caesar) and culinary/herbal Latin (Apicius), same four axes as before:

corpus                              n    wlen     TTR      Hc     bpc
VOYNICH (target)                37554   5.072   0.216   2.341   2.534
Latin: classical (Caesar)       22096   6.128   0.285   3.509   2.788
Latin: culinary (Apicius)        7383   5.837   0.254   3.380   2.450
Latin: alchemical (ps-Llull)    71108   5.794   0.288   3.525   2.803

Hc gap to Voynich (the rigidity tell):
  classical (Caesar)     +1.168
  culinary (Apicius)     +1.039
  alchemical (ps-Llull)  +1.184   <- farthest, not closest

The alchemical Latin is not closer to the Voynich — it's marginally the farthest. All three Latins cluster at conditional entropy Hc ≈ 3.4–3.5; the Voynich sits way down at 2.34, a gap of ~1.0–1.2 bits in every case. Being Lullist, being alchemical, being about secret matter buys you nothing — the rigidity tell is a property of Latin morphology itself, and no register of Latin escapes it. (If anything the Testamentum is the worst match, partly because 16th-c. OCR noise nudges entropy up — but note that cuts for the conclusion: noise raises entropy, so a noisy corpus is a conservative test, and even with the thumb on the scale it can't get within a full bit of the Voynich.)

So the Franciscan → Lullism → alchemical-Latin train terminates here, with three independent negatives stacked up: 1. Llull's wheels (combinatorial generator) overshoot vocabulary and miss the page/grammar structure. 2. Bacon's concealment methods can't lower conditional entropy. 3. Real pseudo-Lullian alchemical Latin carries ordinary Latin entropy, nowhere near the manuscript.

Every road through the Franciscan tradition fails on the same metric — the Voynich's low conditional entropy / high redundancy — which is exactly the signature a generative, self-copying process produces and a meaningful source text (in any language, any cipher, any tradition) does not. The friars were a red herring inherited from the Bacon legend; the evidence keeps pointing at process, not author.

Script voynich_lull_corpus.py; I'll push it to the public repo with the others so this dead end is documented too.

Source: Raymundi Lulli Testamentum, Cologne 1566 (Internet Archive item ARes25626, OCR full text); classical/culinary Latin as in the prior runs.