The Naibbe Key Problem — Why the Voynich Remains Undecoded
A cipher can be explained and still be unreadable. Michael A. Greshko's 2025 Naibbe cipher does not solve the concept voynich manuscript; it makes the failure sharper by showing how Latin or Italian could become Voynich-like text through a hand-executable process using dice, cards, and tables. The hard part is not the mechanism. The hard part is that the missing key may have been a one-time stream of physical events from the early 1400s.
The mechanism
The Voynich Manuscript sits at Yale as Beinecke MS 408: roughly 240 surviving pages, written in an unknown script, on parchment radiocarbon-dated to 1404-1438. Its text behaves oddly. It has word-like structure, repeated glyph patterns, section differences, and a vocabulary shape that has resisted every clean match to a known language or ordinary medieval cipher.
Naibbe is a proposed generator, not a decipherment. It takes Latin or Italian plaintext, uses a die roll to split the stream into single letters and letter pairs, then uses a card draw to choose one of several encoding tables. The name points to naibi, early Italian playing cards recorded in the 14th century.
That matters because the system creates many ciphertext forms for the same plaintext fragment. A short word can explode into multiple surface shapes, which helps explain why Voynichese has so many word-types without behaving like random noise. Greshko's claim is narrow: a 15th-century person could plausibly generate text with several Voynich-like statistics using tools available before printing took over European book production.
The key that disappears
The Naibbe key problem is the gap between producing similar ciphertext and reading the original manuscript. A decoder would need the table mappings, the exact sequence of die outcomes, the exact sequence of card draws, and the plaintext language. None is known.
That is different from a locked box. A lost password might exist in a letter, a memory, or a reused system. A die roll from 1420 does not exist anywhere after the throw. If a scribe used fresh random choices across the manuscript and did not preserve the run sheet, the key is not hidden. It is gone.
Shannon's 1949 "Communication Theory of Secrecy Systems" gives the clean frame. A cipher breaks when the redundancy of the message overwhelms uncertainty in the key. Enigma traffic had cribs: weather formats, repeated military phrasing, operator habits. Voynich has no confirmed crib. No plant name, zodiac label, recipe heading, or Latin phrase has earned consensus.
| Case | Known mechanism? | Known crib? | Key status | Result |
|---|---|---|---|---|
| Enigma traffic | Yes | Often | Daily settings recoverable | Broken at scale |
| One-time pad | Yes | Not enough | Random pad destroyed | Secure if used correctly |
| Voynich under Naibbe-like encoding | Maybe | No | Tables and random stream missing | No convergence |
What's contested
The live question is not "is Naibbe the answer?" Greshko himself does not claim that. The live question is whether a Naibbe-like family of verbose homophonic ciphers can explain more of Voynichese than language, hoax, or generated pseudo-text models.
Lisa Fagin Davis's 2020 paleographic work argues for five scribes, which weakens the lazy version of the single-forger story. It does not prove meaningful content. Five people can copy nonsense under shared rules, just as one person can encode meaningful text with private rules.
The strongest point against easy decipherment is empirical: more than a century after Wilfrid Voynich bought the manuscript in 1912, no proposed reading has produced a repeatable grammar, a stable lexicon, and independent confirmation from the illustrations. The manuscript keeps giving structure without giving reference.
Why this has to do with other realms
The Naibbe key problem belongs to cryptography, but it also smells like physics. A key made of dice and cards is a historical event stream. Once the dice stop moving and nobody records the outcome, the information has no public trace to recover.
That is why the bridge to concept information theory is stronger than the bridge to ordinary puzzle-solving. The manuscript asks whether meaningful information can be present in an artifact while the route back to meaning has been physically erased. The same boundary appears in concept one time pad, in archaeological reconstruction, and in any system where the measurement record matters more than the object left behind.
An open question
If the Voynich text is meaningful, what kind of external anchor could still be strong enough to start convergence: one plant, one star chart, one recipe, or one marginal note that ties a glyph sequence to a known word?
Key Sources
- Michael A. Greshko, "The Naibbe cipher: a substitution cipher that encrypts Latin and Italian as Voynich Manuscript-like ciphertext," Cryptologia, 2025 - the page's central modern cipher proposal.
- Claude E. Shannon, "Communication Theory of Secrecy Systems," Bell System Technical Journal, 1949 - the key entropy and redundancy frame.
- Lisa Fagin Davis, "How Many Glyphs and How Many Scribes? Digital Paleography and the Voynich Manuscript," Manuscript Studies, 2020 - the five-scribe evidence.
- Claire Bowern and Luke Lindemann, "The Linguistics of the Voynich Manuscript," Annual Review of Linguistics, 2021 - survey of linguistic structure and failed hypotheses.
- Sravana Reddy and Kevin Knight, "What We Know About the Voynich Manuscript," ACL Workshop, 2011 - computational baseline for language and writing-system tests.
- Rachel Sterneck, Annie Polish, and Claire Bowern, "Topic Modeling in the Voynich Manuscript," arXiv:2107.02858, 2021 - topic clusters compared with scribes and illustrations.
Further Reading
- Yale Beinecke MS 408 digital facsimile - the manuscript itself, not screenshots filtered through someone else's theory.
- M. E. D'Imperio, The Voynich Manuscript: An Elegant Enigma, NSA, 1978 - still useful because it catalogues old claims without pretending they solved the text.
- René Zandbergen's Voynich Manuscript site - a careful archive of provenance, imagery, and transcription issues.
- concept enigma machine - the counterexample where known cribs turned machine encryption into a daily labor problem.
- concept beale ciphers - a useful contrast: a cipher where the suspected key is a public text rather than a vanished event stream.
See Also
- concept voynich manuscript
- concept information theory
- concept one time pad
- concept enigma machine
- concept beale ciphers
- concept william friedman voynich
- concept probability cipher connection
Abhishek's take
The useful part of Naibbe is that it lowers the drama and raises the bar. I do not need the manuscript to be mystical, alien, or fraudulent; the stranger possibility is that it may be ordinary plaintext behind a one-time analog procedure nobody recorded. That is a better warning for builders than a puzzle story: a system can preserve output for 600 years and still lose the one thing needed to read it.
Tags: #voynich #cryptography #information-theory #cipher #entropy #key-recovery #naibbe