The origin of the Voynich MS and of its stochastic text—finally revealed through deep search of historical records
“G. A.”, 1 April 2026.
Abstract
An AI-assisted deep search of historical records has yielded conclusive new information concerning the origin of the Voynich MS. This evidence requires us to rethink everything we thought we knew about the manuscript. This study will demonstrate that two unlikely theories about the manuscript, when combined, create a coherent and fully consistent explanation for both the manuscript’s origin, and the nature of its text.
1. Introduction
The Voynich MS is generally believed to be a late medieval work, most likely produced in Central Europe. Its fame rests on its mysterious text, widely assumed to be a cipher, yet resistant to all attempts at decryption. This presents a remarkable mystery: even the most advanced modern cryptanalytic techniques have failed to make meaningful progress on a text over six centuries old.
Over the last few decades, numerous theories have been proposed, ranging from specific decipherments to broader interpretations of the manuscript’s purpose. None, however, have been proven convincing. In this study, two of the more controversial hypotheses will take the centre stage.
The first is the proposal that the manuscript is not medieval, but a modern forgery ([Barlow], [Santacoloma]). This idea has attracted little support, largely because it lacks direct evidence and fails to address the central problem of the unreadable text.
The second is the so-called auto-copy hypothesis ([Timm]), which proposes a mechanism by which a meaningless text could be generated rapidly, while reproducing some of the statistical features observed in the Voynich MS text. This theory, too, has found little traction, both due to lack of evidence and because it appears anachronistic in a medieval context.
By just putting these two hypotheses side by side, it immediately appears that they are capable of compensating each other’s weaknesses: the modern forgery hypothesis resolves the anachronism of the auto-copy hypothesis, and the quick-and-easy text generation method seems to be just what the modern forgery hypothesis was lacking.
This is not just another hypothesis. An AI-assisted deep search of historical records has uncovered significant new evidence that demonstrates that this is exactly what happened.
2. AI-assisted deep search
A specially configured AI system was supplied with extensive data concerning the Voynich MS—its text, history, and prior analyses—and tasked with identifying hidden patterns and unexplored connections. This initial phase, begun on 4 August 2025, may be termed the first wave.
After several iterations, a striking pattern emerged: all significant leads pointed to St. Petersburg, and more broadly to Russia. Notably, these connections appeared only for events after 1773, with no relevant links identified before that date. Although this temporal cutoff initially cast doubt on the findings, the absence of alternative leads justified a second, more targeted phase of investigation.
During this second wave, previously unknown digital archives, relating to international intelligence activities, were identified. These materials—referred to here as the “S-07 archive”—contained a number of remarkable details. Although the archive disappeared from the internet shortly after its discovery, substantial information was retrieved in time. Importantly, several specific details found in the S-07 archive could be independently corroborated through modern sources, as will be indicated in the following.
Four principal lines of inquiry emerged in the first wave and were explored more deeply in the second wave. These are:
- The suppression of the Society of Jesus in 1773, and the subsequent refuge offered to Jesuits in Russia by Catherine the Great
- The work of Andrey Andreyevich Markov in St. Petersburg between 1886 and 1910, particularly on recursive text generation
- The presence of Ethel Lilian Voynich in St. Petersburg shortly before 1900, and her connections to anti-Tsarist groups
- Intelligence activities between 1890 and 1925, involving a Russian agent initially operating in Britain, and later in the US.
3. Roman Jesuits in Russia
The Jesuit father Jacob Pierling (1784 - 1870) was born in St. Petersburg and later became one of the leading figures in the Society of Jesus, among others as the father provincial of Austria, during the time of restoration. He recalls in his memoirs that, during his childhood, he was instructed by Roman Jesuits, who had found refuge in Russia after the suppression of the Society of Jesus in most of Western Europe and their colonies. He learned from them that they had also transferred parts of their most valuable historical documents to Russia for safe keeping. The extent of this material was not recorded by Pierling and could not be discovered in the S-07 archive. It is also not known whether this material was later returned to Rome, remained in Russia, or must now be considered lost.
4. Markov text generation
The mathematician Andrey A. Markov (1857-1922), a professor at the University of St. Petersburg, is best known for developing what are now known as Markov processes or Markov chains—stochastic systems in which each element depends probabilistically on preceding elements.
Markov’s interest extended to the statistical structure of language, particularly at the level of characters, independent of meaning. In 1908, following political unrest, he was ordered by the government to monitor his students. His refusal led to his resignation. To avoid harsher consequences, he agreed to undertake a government-requested project: to apply his recursive processes to the generation of a text that appeared meaningful while in fact being devoid of semantic content.
Markov never knew the purpose of this activity, but he appears to have obliged without asking questions. According to the S-07 archive, the method consisted of setting up a chain of words by applying probabilistic changes to earlier words in the chain. A concern was raised that the method was too easy to detect from the resulting text, and this was fixed by mixing the outputs from sometimes two, sometimes three separate recursive chains, generated with different settings. Markov produced his text using the Russian alphabet of 36 characters, but this was later converted into an invented script.
5. Ethel Lilian Boole/Voynich
Ethel Lilian Voynich is widely known as a novelist and as the wife of Wilfrid Voynich, but her earlier years in Russia have attracted less interest. While studying in Berlin in 1885, she developed a strong interest in Russian revolutionary movements and quickly mastered the language. In 1887 she travelled to St. Petersburg, where she stayed with the Karauloff family, who were relatives of the revolutionary Sergey Stepnyak-Kravchinsky. The latter had already fled the country to live in London.
She soon became known as a strong anti-Tzarist, and she was approached by agents of Russian intelligence posing as fellow revolutionaries. They persuaded her to assist a future Russian dissident who would also flee to London. This man would identify himself using a code phrase. In May 1889, she left Russia, smuggling a manuscript for Stepniak. Although this document has occasionally been linked to the Voynich Manuscript, that is not the case, as we shall see in the following.
6. The Russian agent—early years
The Russian agent mentioned above is none other than Wilfrid Voynich, or, more precisely, a person assuming Voynich’s identity. A summary biography combining the two lives is given by [Zandbergen]. The real Voynich, detained in a prison camp in Tunka, Siberia, died while escaping from Irkutsk on 12 June 1890. His identity was then adopted by an agent pretending to be an anti-Tzarist revolutionary.
He travelled to London using forged documents, where he integrated with other revolutionary networks, to gain trust in preparation for his future tasks. Interestingly, a glitch occurred as he arrived already on 5 October 1890, which is faster than any ship could have taken him there. This discrepancy was already pointed out by [Prinke]. The S-07 archive provides the interesting detail that the code phrase to identify himself to Ethel Boole was: “I saw you from the window of my prison cell in the Warsaw citadel”. This phrase had so far puzzled biographers.
Ethel and pseudo-Wilfrid (simply Wilfrid in the following), became a dedicated couple. They came from completely different backgrounds, but their dedication to their cause kept them together. Only Ethel did not know that Wilfrid’s real cause was a different one. In the early years they were regularly followed by Russian agents. [Kennedy] reports an interesting anecdote, which shows that they were well aware of this, and they would make it a game to spot these agents and expose them. But of course, these agents presented no danger.
Beside participating in revolutionary groups, Wilfrid also needed a regular occupation, and he became a rare book antiquarian. This had the added advantage that he could travel extensively without raising suspicion.
7. The mystery book and later years
The S-07 archive indicates that after 1900 activities in Britain were slowing down, and interest was focussing more on the US, where cryptologic intelligence services were undergoing a major change ([Furlong]). Systems based on code books were gradually being replaced by more advanced rotor systems. In order to disturb this progress, it was decided to present these agencies with an unsolvable cipher, in a way that would not raise any suspicion. It was for this reason that the cipher manuscript was devised, based on Markov’s recursive, meaningless text, generated in 1908.
The book was completed in 1909 and delivered to Voynich through the book dealer Joseph Baer of Frankfurt. Here, another glitch occurred. It was intended that the book should only appear after a sale of manuscripts from the Society of Jesus, but the book was accidentally seen by the German botanist Charles Singer, already a few years before this sale took place.
As planned, the book was never offered for sale in Britain, and when Voynich moved to the US in 1914, it was also not for sale there. However, by this time, Voynich had acquired a fortune from his book sales to the extent that his loyalty to the Russian cause became a reason for concern. For that reason, Ethel was relieved of her duty, and a new agent was sent to accompany Wilfrid, to keep him under a closer watch. This was the very capable and very resolute Anne Margret Nill. Her control over Voynich was noticed by his former assistant E. Millicent Sowerby, who recorded this in her memoirs ([Sowerby]).
Once Voynich was in the US, he made every effort to interest military codebreaking agencies into its decryption. By all appearances, the original purpose of the MS seems not to have been achieved. While the text was the subject of significant attention by cryptologists, this was mainly done during their spare time.
8. Conclusion
This paper demonstrates beyond reasonable doubt that the Voynich MS is a modern fake conceived in St. Petersburg around the start of the 20th century. Contrary to earlier ‘modern fake’ theories, the MS was not created by Wilfrid Voynich.
The text of the MS was deliberately designed as an unsolvable cipher, using a recursive method conceived in 1908, similar to the so-called auto-copy method proposed in recent years.
The purpose of this MS was to confuse and delay American cryptological services. As far as can be concluded, this was not achieved.
Acknowledgments
This paper claims to have been generated with significant support from AI.
Bibliography
Barlow, M.: The Voynich Manuscript --- by Voynich?, 1986.
Furlong, Herbert: The birth (1887 - 1900) and dawn (1900 – 1917) of US cryptology, 1938.
Kennedy, Gerry: The Booles & The Hintons, two dynasties that helped shape the modern world, 2016.
Ма́рков, Андре́й: Эксперименты по генерации бессмысленного текста с использованием стохастических рекурсивных методов, 1908
Pierling, Jakob: Livres rares et lettres inédites des Jésuites Romains en Russie, après le décret ‘Dominus Redemptor’ de 1773, 1812.
Prinke, Rafał: History of the Manuscript from Rudolf to Voynich, 2012.
Santacoloma, Rich: The 1910 Voynich theory (You are not allowed to view links.
Register or
Login to view.), 2012
Sowerby, E. Millicent: Rare People and Rare Books, 1967.
Taratuta, Evgeniya: our friend Ethel Lilian Voynich/Boole, 1957 (translated by Seamus O Coigligh, 2008).
Timm, Torsten and Andreas Schinner: A possible generating algorithm of the Voynich manuscript, 2019.
Zandbergen, René: Biography of Wilfrid Voynich (You are not allowed to view links.
Register or
Login to view.), 2025.