A Workshop Cipher: A Generative Model Reproducing the Statics and Dynamics of the Voynich Manuscript Text

{"The":[0,240,341],"text":[1,268],"of":[2,51,93,115,125,167,226],"the":[3,52,79,83,86,90,154,159,171,173,184,200,219,248,255,267,324],"Voynich":[4],"Manuscript":[5],"(Beinecke":[6],"MS":[7],"408)":[8],"combines":[9],"language-like":[10],"statistics":[11],"(Zipf":[12],"law,":[13],"vowel-consonant":[14],"alternation,":[15],"topic":[16],"structure)":[17],"with":[18,104,122,301,322],"properties":[19],"no":[20],"natural":[21],"language":[22],"shows,":[23],"most":[24],"notably":[25],"a":[26,43,67,74,108,131,189,243,246],"conditional":[27],"character":[28],"entropy":[29],"h2":[30],"~":[31,129,146],"2.1-2.3,":[32],"far":[33],"below":[34],"all":[35,98],"tested":[36],"languages.":[37],"Greshko":[38],"(Cryptologia,":[39],"2025)":[40],"showed":[41],"that":[42,62],"verbose":[44],"homophonic":[45],"table":[46,168,236],"cipher":[47,110],"(\\"Naibbe\\")":[48],"reproduces":[49,82],"much":[50],"static":[53,84,220],"profile;":[54],"Parisel":[55,100],"(2026)":[56],"introduced":[57],"four":[58,99,105],"dynamic":[59],"word-boundary":[60],"signatures":[61,102],"his":[63],"own":[64],"generators":[65],"and":[66,97,141,188,195,230,270,285,304,308,310,320],"Cardan":[68],"grille":[69],"fail":[70],"jointly.":[71],"We":[72],"present":[73],"generative":[75],"model":[76,241,256],"that,":[77],"for":[78],"first":[80],"time,":[81],"profile,":[85],"homophone":[87],"\\"twin\\"":[88],"layer,":[89],"serial":[91,123,193],"clustering":[92,194],"homophones":[94],"within":[95],"lines,":[96],"boundary":[101,174],"simultaneously,":[103],"interpretable":[106],"parameters:":[107],"Naibbe-class":[109],"(six":[111],"balanced":[112],"tables,":[113],"units":[114,127],"one":[116],"or":[117],"two":[118],"plaintext":[119],"letters)":[120],"operated":[121],"reuse":[124],"recent":[126],"(lambda":[128],"0.22),":[130],"weak":[132],"graphotactic":[133],"choice":[134,169],"at":[135,204,288],"word":[136],"boundaries":[137],"(K":[138],"=":[139],"2),":[140],"moderate":[142],"spelling":[143],"noise":[144],"(epsilon":[145],"0.1).":[147],"Homophone":[148],"twin":[149,196,206],"distributions":[150],"are":[151],"invariant":[152],"across":[153],"five":[155,225],"scribal":[156],"hands,":[157],"identifying":[158],"apparatus":[160],"as":[161],"shared":[162],"workshop":[163],"technology;":[164],"any":[165],"skew":[166],"degrades":[170],"fit;":[172],"preferences":[175],"can":[176,269],"be":[177],"derived":[178],"from":[179,234],"intra-word":[180],"transition":[181],"statistics,":[182],"removing":[183],"last":[185],"circular":[186],"element;":[187],"plaintext-genre":[190],"gradient":[191],"moves":[192],"rate":[197,207],"together,":[198],"implying":[199],"manuscript's":[201],"high":[202],"seriality":[203],"low":[205],"is":[208,242,250,257,283,344],"procedural,":[209],"not":[210,238,245,251],"textual.":[211],"A":[212],"self-citation":[213],"(\\"meaningless":[214],"text\\")":[215],"generator":[216],"fitted":[217],"to":[218],"profile":[221],"remains":[222],"separable":[223],"on":[224,263],"six":[227],"preregistered":[228],"windows,":[229],"long-range":[231],"burstiness":[232],"arises":[233],"section-level":[235],"dialects,":[237],"topicality.":[239],"mechanism,":[244],"decipherment:":[247],"key":[249],"recovered,":[252],"even":[253],"when":[254],"inverted":[258],"before":[259],"attack;":[260],"we":[261],"argue":[262],"capacity":[264],"grounds":[265],"what":[266],"cannot":[271],"encode.":[272],"Replication":[273,296],"package":[274,297,325],"(87":[275],"scripts,":[276],"corpora,":[277],"labels":[278],"dataset,":[279],"CLAIMS,":[280],"README":[281],"EN/RU)":[282],"attached":[284],"also":[286],"available":[287],"https://voynich.site/files/voynich_replication_v1.zip":[289],".":[290],"Project":[291],"site:":[292],"https://voynich.site":[293],"v1.1":[294],"(2026-09-21).":[295],"updated:":[298],"added":[299],"test_parisel4.py":[300],"its":[302],"reports":[303],"pre-registration":[305],"(test":[306],"#61),":[307],"analysis/external/seeton":[309],"analysis/external/vgcmi":[311],"-":[312],"third-party":[313],"modules":[314],"imported":[315],"by":[316,338],"packaged":[317],"scripts":[318],"(Apache-2.0":[319],"MIT,":[321],"notices);":[323],"assembler":[326],"now":[327],"follows":[328],"imports":[329],"transitively.":[330],"333":[331],"files,":[332],"3.3":[333],"MB,":[334],"sha256":[335],"c01688ebd93e9f6dc99a06dee26f13e27f71054c6672a0cbcc43af32f2865fc2.":[336],"Reported":[337],"Igor":[339],"Golub.":[340],"paper":[342],"PDF":[343],"unchanged.":[345]}

Authors

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-09-21
DOI
https://doi.org/10.5281/zenodo.22874156
Primary Topic
Intelligence, Security, War Strategy
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
preprint

A Workshop Cipher: A Generative Model Reproducing the Statics and Dynamics of the Voynich Manuscript Text

Vitaly Averyanov
Zenodo (CERN European Organization for Nuclear Research)
Intelligence, Security, War Strategy
preprint

A Workshop Cipher: A Generative Model Reproducing the Statics and Dynamics of the Voynich Manuscript Text

Vitaly Averyanov
preprint en

Abstract

The text of the Voynich Manuscript (Beinecke MS 408) combines language-like statistics (Zipf law, vowel-consonant alternation, topic structure) with properties no natural language shows, most notably a conditional character entropy h2 ~ 2.1-2.3, far below all tested languages. Greshko (Cryptologia, 2025) showed that a verbose homophonic table cipher ("Naibbe") reproduces much of the static profile; Parisel (2026) introduced four dynamic word-boundary signatures that his own generators and a Cardan grille fail jointly. We present a generative model that, for the first time, reproduces the static profile, the homophone "twin" layer, the serial clustering of homophones within lines, and all four Parisel boundary signatures simultaneously, with four interpretable parameters: a Naibbe-class cipher (six balanced tables, units of one or two plaintext letters) operated with serial reuse of recent units (lambda ~ 0.22), a weak graphotactic choice at word boundaries (K = 2), and moderate spelling noise (epsilon ~ 0.1). Homophone twin distributions are invariant across the five scribal hands, identifying the apparatus as shared workshop technology; any skew of table choice degrades the fit; the boundary preferences can be derived from intra-word transition statistics, removing the last circular element; and a plaintext-genre gradient moves serial clustering and twin rate together, implying the manuscript's high seriality at low twin rate is procedural, not textual. A self-citation ("meaningless text") generator fitted to the static profile remains separable on five of six preregistered windows, and long-range burstiness arises from section-level table dialects, not topicality. The model is a mechanism, not a decipherment: the key is not recovered, even when the model is inverted before attack; we argue on capacity grounds what the text can and cannot encode. Replication package (87 scripts, corpora, labels dataset, CLAIMS, README EN/RU) is attached and also available at https://voynich.site/files/voynich_replication_v1.zip . Project site: https://voynich.site v1.1 (2026-09-21). Replication package updated: added test_parisel4.py with its reports and pre-registration (test #61), and analysis/external/seeton and analysis/external/vgcmi - third-party modules imported by packaged scripts (Apache-2.0 and MIT, with notices); the package assembler now follows imports transitively. 333 files, 3.3 MB, sha256 c01688ebd93e9f6dc99a06dee26f13e27f71054c6672a0cbcc43af32f2865fc2. Reported by Igor Golub. The paper PDF is unchanged.

Zenodo (CERN European Organization for Nuclear Research)
Quality Education
Intelligence, Security, War Strategy
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.