The Rohonc Codex, one of the world’s most famous crypto mysteries, appears to have been solved. Here are the details about the solution.

Part 1

Here is the second part of my FAQ about the recently published solution of the Rohonc Codex.


Who has deciphered the Rohonc Codex?

The solution in question was published by the Hungarian scientists Levente Zoltán Király and Gábor Tokai in the scientific magazine Cryptologia (the following photographs were provided by Levente Zoltán Király).

Zoltán Király …


… was born in 1976. He earned a BSc in computer science and an MA in divinites. He is currently working as a software developer. He has occupied himself with the Rohonc Codex as a hobby since 2005.

Gábor Tokai, …


… born in 1969, is an archaeologist and art historian. He has worked at the Hungarian National Gallery in Budapest since 1999. Tokai has dealt with undeciphered scripts from the age of sixteen.

Király and Tokai started their research on the Rohonc Codex independently from each other. The two met at a conference in late 2010, where Király summarized his results in a lecture. Having acknowledged common results that they had reached independently, they decided to work together. The following picture was taken in Rechnitz (Hungarian: Rohonc), Austria, the town the codex is named for.



What kind of encryption method was used?

According to Király and Tokai’s paper, the text in the Rohonc Codex is encrypted in a code. This means that every word in the ciphertext of the codex stands for a certain cleartext word. To create the text, the author of the Rohonc Codex had to use a (self-developed?) codebook. If Király and Tokai’s claim is correct, it is not possible to decipher the Rohonc Codex based on a letter substitution table.

Codes (especially nomenclators) were quite common in the 16th century. It is therefore not surprising to encounter a code-based encryption in a 16th century document. However, I have never heard of a complete book encrypted this way, let alone one that is 500 years old.

Király and Tokai’s paper doesn’t mention the language the cleartext is written in. They plan to address this topic in a future publication. It is generally possible to read a message written in code without knowing the cleartext language, as long as the meanings of the code words are known.

In addition, Király and Tokai don’t mention if the author of the codex used the code for secrecy or for another reason. I hope, future publications of the two will address this question, too.


What is the content of the Rohonc Codex?

As the pictures (most of them show scenes from the bible) suggest, the content of the Rohonc Codex is Christian in nature. Király and Tokai have found several prayers, Bible stories, and Bible quotes. Most of the text refers to the New Testament. It is an interesting, yet unanswered question why a text of this kind (as far as can be told so far, it doesn’t contain any secrets) was encoded in such a way.


How did Király and Tokai decipher the Rohonc Codex?

Already before Király and Tokai started their work, the two researchers Ottó Gyürk and Miklós Locsmándi had identified numbers in the Rohonc Codex. These numbers follow a system similar (but not equal) to the Roman numbers. The one of this number system is represented by a vertical line. In the following picture seven ones in a row can be seen.


Apart from the numbers, Király and Tokai used words written in the pictures as cribs.


Finally, the codex’s literal quotations from the Bible proved helpful for breaking the ciphertext. Király and Tokai consider these quotations as a kind of Rosetta Stones of the Rohonc Codex. For instance, knowing the code word for “give” and having a conjecture about the code word for “tree”, Tokai could identify a passage in the codex as Matthew 7:17–18.

Based on this findings, Király and Tokai could decipher considerable parts of the Rohonc Codex.


Is the Rohonc Codex completely deciphered now?

No. Király and Tokai make clear that the present state of their research project is not marked by finished research. They say that it is likely that the code will never be completely broken (this is not uncommon, as the meaning of some code words cannot be derived from the context).

Király and Tokai now hope for help from specialists of other disciplines—such as linguistics or historians. This is why they have decided to publish their findings the way they currently are.

Király and Tokai plan to publish further articles, in which they plan to describe the content of the cleartext, as well as the morphology, syntax, and peculiarities of the code. The possible language(s) of the author will be addressed, too. They will also provide a detailed table of contents, further indices, and a glossary.


Is Király and Tokai’s solution correct?

To be honest, it is hard for me to judge whether Király and Tokai’s solution is correct. I am a crypto expert, not a linguist. However, I know that Cryptologia has high proof-reading standards (I am a member of the magazine’s editorial board and I occasionally do proof-readings) and that Craig Bauer, the editor-in-chief of Cryptologia, is an experienced cryptogram expert who would certainly not publish a non-substantial solution.

In addition, I know that Rohonc Codex expert Benedek Láng has followed the path of Király and Tokai’s solution in the last 8 years and that he has read the article very carefully. He thinks that the solution is correct.

All in all, I put a fair degree of trust in this solution.

I hope that future publications of Király and Tokai will reveal more details about the solution and that more experts will state their opinion about it.

Further reading: How Paolo Bonavoglia solved a 19th century encrypted book


Subscribe to Blog via Email

Enter your email address to subscribe to this blog and receive notifications of new posts by email.

Kommentare (7)

  1. #1 Rich SantaColoma
    5. Juni 2018

    “… the Rohonc Codex is encrypted in a code. This means that every word in the ciphertext of the codex stands for a certain cleartext word. To create the text, the author of the Rohonc Codex had to use a (self-developed?) codebook. If Király and Tokai’s claim is correct, it is not possible to decipher the Rohonc Codex based on a letter substitution table.”

    I would wonder, though, if there is still the possibility that this is not a code, but possibly a cipher, or constructed language, using syllabic signs. As an example of what I mean, say a person was writing a language and used a symbol of a slice of pie to denote the sound for “piece”, while meaning “peace”. This sort of thing has been done in some Mesoamerican writing, and in Linear B, and in Egyptian Hieroglyphics, and others. That is, the symbol for an object is used for the sound of that object in the spoken language, but NOT for the object itself.

    In this case, for instance, the decoders have determined a certain character set means “peace”. But rather than being an arbitrary code symbol denoting “peace”, perhaps there is, in some language, a phonetic similarity between the symbol drawn, and the word “peace” in that language.

    This would not counter the elements of this translation attempt. But perhaps the idea could help both uncover the underlaying language, and at the same time, then, help decipher the rest of the work. Well, that is, if it is a syllabary, and not a code as suspected.

  2. #2 Davidsch
    6. Juni 2018

    @Rich, a phonetic or partial phonetic iconographic language is possible and I think I can check if the text is similar to a language.

    However I have never looked into the transcript of this text, if I have time I will give it a go.

    The above proposed nomenclator solution is nonsense if there is no original cipher book. Any child can make up his or her own fantasy text, therefore I already reject this solution.

  3. #3 Davidsch
    6. Juni 2018

    ps. Is there a transcript and font of the (original) text available?

  4. #4 Klaus Schmeh
    6. Juni 2018

    >Is there a transcript and font of
    >the (original) text available?
    As far as I know, this is not the case.

  5. #5 Torsten Timm
    6. Juni 2018

    @Klaus The authors argue “To give weight to our argument we present several interlinear translations of passages from the codex`s text.”
    But how can we check the suggested interlinear translation? The Rohonc Codex didn’t use any spaces. This means, that we do not know if symbols belong together or not. Therefore we can’t use the word spaces to check any decipherment attempt. Another problem is that the language behind the codex is still unidentified. With other words it is also not possible to check if the given word order is following the rules of a language or not. Last but not least Király and Tokai claim that there decipherment attempt is based on some Bibel quotations. Therefore it is no surprise if some parts of the suggested decipherment sound like Bibel quotations. Therefore the question is still open in my eyes: How can we check the validity of the suggested decipherment?

  6. #6 Charlotte Auer
    7. Juni 2018

    Just to underline my previous doubts:

    “Finally, the codex’s literal quotations from the Bible proved helpful for breaking the ciphertext.”

    How can someone identify “literal” quotations from the Bible without knowing the underlying language of the codex? Furthermore, in the early 16th Century was no kind of common standard in biblical texts existing, but a lot of exegeses and variations. So literal quotations would stringently require the quoted Bible itself to compare with.

    To me that attempt of the decipherment looks more than arbitrary.

  7. #7 Davidsch
    25. September 2019

    After careful study of the papers, books and the manuscript, I conclude this cipher is NOT solved at all.

    The proposed “solution” (publication Kiraly et al, 1 from 3) is a vague attempt where ideograms were used and “identified”. Then based on those assumptions a storyline from a gospel was chosen by the “solvers”. LInes from that gospel were searched and chosen to make sentences based on ideograms that could be logical. Thus, a couple of lines perhaps one or two pages were “done”. I do not accept that as an solution.

    I started now a transcript myself, because nobody wants to share their work already done, but since it is very tedious, it will take me at least a whole year. Based on such full transcript, analysis could be made.