One of the most persistent shortcuts in manuscript localisation is linguistic.
If a text looks Italian, the manuscript must be Italian.
If a plant name is German, the manuscript must be German.
If a technical term is Latin, the manuscript belongs to learned university culture.
Medieval technical manuscripts make all three inferences dangerous.
Language travels with knowledge. A manuscript can be physically produced in one place, written in several languages, preserve vocabulary from another region and transmit a technical tradition whose intellectual ancestry is somewhere else again.
Quick Read
- Technical manuscripts frequently mix Latin and vernacular languages.
- Salzburg M I 36, produced at Laubach in 1425, contains German and Latin technical material and a plant catalogue using Italian, Latin and German.
- Klosterneuburg Cod. 684, from the 1410s, combines astronomical and medical texts in Latin and German.
- Later fifteenth-century technical manuscripts from Naples studied by modern scholarship can mix Latin, Catalan and Italian recipes inside one compilation.
- These examples show that language, source tradition and production geography can diverge sharply.
- Therefore a future Voynich decipherment yielding Italian, German, Latin or mixed vocabulary would not by itself identify where the physical manuscript was made.
- The geographic value would depend on local forms, scribal habits, dialect, technical terminology and independent object evidence.
A Technical Manuscript Is a Meeting Place
Technical knowledge is unusually mobile.
A remedy can move from Arabic into Latin.
Latin can become Italian.
An Italian plant name can travel north.
A German practitioner can copy a Latin authority.
A recipe can preserve one language for the ingredient and another for the instruction.
The manuscript becomes a contact zone.
technical transmission rewards usefulness more than linguistic purity.
Laubach 1425 Is the Strongest Exact-Window Language Control
Salzburg University Library M I 36 was produced in Laubach/Wetterau in 1425.
Its broad technical miscellany includes German and Latin material across medicine, astrology, astronomy, prognostics and practical knowledge.
Most strikingly, its plant catalogue includes Italian, Latin and German.
This is direct evidence that Italian botanical vocabulary can appear inside a German-produced technical manuscript inside the Voynich parchment window.
That single fact should permanently weaken one simplistic localisation rule:
Italian word ≠ Italian production.
Klosterneuburg Shows Latin and German Coexisting in the 1410s
Klosterneuburg Cod. 684 is catalogued as a composite manuscript of astronomical and medical texts from the second decade of the fifteenth century.
Its languages include Latin and German.
This matters because the manuscript sits very close to the start of the Voynich material window.
A mixed technical-language environment is therefore not a late development.
It belongs to the exact period.
The Carrara Herbal Gives the Reverse Transformation
The Carrara Herbal provides a different language route.
Imported materia medica knowledge associated with Arabic and Latin traditions is rendered into a Paduan vernacular manuscript.
This demonstrates:
foreign knowledge → Latin mediation → local vernacular representation.
Laubach demonstrates the reverse kind of complexity:
German production environment → multilingual technical vocabulary including Italian.
Together they show why language and geography must be separated.
A Successful Voynich Decipherment Could Still Be Geographically Ambiguous
Imagine tomorrow that Voynichese is convincingly shown to encode an Italian vernacular.
That would be a major breakthrough.
It would not automatically prove Paduan production.
An Italian-speaking or Italian-reading practitioner could write elsewhere.
An Italian source could be copied in Prague.
A German student could bring Paduan notes home.
A Venetian physician could use Paduan terminology.
Language would narrow intellectual and social geography more than physical geography.
Dialect Could Carry More Geographic Weight
Not all linguistic evidence is equally weak.
A generic Italian word travels easily.
A distinctive local dialect form may travel less widely.
A cluster of Paduan vernacular forms, especially in low-prestige function words or routine morphology, could carry stronger geographic signal.
Even then, the signal must be compared against:
- Venetian;
- Veronese;
- Lombard;
- Emilian;
- transalpine copies of Italian texts.
One exotic plant name is weak.
A grammatical system is stronger.
Technical Vocabulary Is Especially Mobile
Medicine, astronomy and pharmacy create travelling specialist vocabularies.
Ingredient names move through trade.
Astronomical terms move through textbooks.
Measures move through professional practice.
Therefore technical nouns may carry less local signal than ordinary grammar.
This is a major warning for crib-based Voynich decipherment.
Recognising one plausible drug name says little about production geography unless the surrounding language behaves consistently.
Code-Switching Is a Real Historical Possibility
A technical compiler can switch language by task.
- Latin for authority.
- Vernacular for execution.
- Foreign plant name for ingredient identity.
- Local gloss for explanation.
- Special symbols for quantity or operation.
Modern scholarship on late-fifteenth-century technical compilations has documented precisely this kind of multilingual entanglement, including Latin, Catalan and Italian within alchemical and recipe collections.
Voynich may or may not use multiple languages.
But multilingual technical culture is historically normal enough to remain in the model set.
An Invented Script Can Hide Multiple Language Layers
The Voynich script adds a second transformation layer.
Even if the underlying language were multilingual, a common invented script could visually normalise it.
That could make different languages look like one surface system.
Conversely, Currier-style statistical differences might reflect function rather than language.
These alternatives must be separated statistically.
Language Mixing Produces Testable Predictions
If different Voynich sections encode different languages, we might expect:
- different word-length distributions;
- different function-word populations;
- different morphological productivity;
- different label/prose relationships;
- hand-independent language clustering.
If differences instead track scribal hands or page functions, a multi-language model weakens.
The Padua Language Hypothesis Needs a Hierarchy
If a future reading points toward northern Italy, claims should be ranked carefully.
- Italian-related language: broad linguistic ancestry.
- Northern-Italian linguistic features: narrower.
- Veneto features: narrower again.
- Paduan vernacular features: stronger regional signal.
- Paduan linguistic features plus Paduan scribal/codicological evidence: meaningful convergence.
Skipping from Level 1 to Level 5 would be another provenance shortcut.
Language Can Also Reveal Routes
Mixed terminology can sometimes reveal transmission history better than production place.
An Italian technical term embedded in German grammar may indicate imported knowledge.
An Arabic-derived drug name in Italian may reveal deeper textual ancestry.
Latin headings with vernacular recipes may reveal receiver change.
These patterns can reconstruct routes.
That may be exactly what Voynich needs.
The Multilingual Test
- Separate scribal hand from textual population.
- Test whether Currier differences persist within the same hand.
- Compare labels and running text separately.
- Search for function-word-like high-frequency tokens before technical nouns.
- Test whether morphology changes by section or by scribe.
- Compare any proposed lexical identifications against multilingual technical controls.
- Use dialect only after a broad reading succeeds across unseen text.
- Keep production geography independent until object evidence converges.
What Multilingual Controls Can Legitimately Support
- Technical manuscripts can mix languages inside one codex.
- Italian vocabulary can occur in a German-produced manuscript.
- Language and production geography can diverge.
- Specialist vocabulary may be especially mobile.
- A decipherment can reveal intellectual route without proving physical origin.
What They Cannot Support
- that Voynich is multilingual;
- that Currier A and B are different languages;
- that Italian words would prove Padua;
- that German-looking forms would prove Central-European production;
- that multilinguality explains the invented script.
Research Sources
- Handschriftencensus — Salzburg M I 36
- manuscripta.at — Klosterneuburg Cod. 684
- M. Del Savio — Multilingual Entanglement in Technical Texts
- British Library — Carrara Herbal, Egerton MS 2020
Continue the Series
- Voynich and Padua — The Script-Geography Problem
- Voynich and Padua — Laubach 1425
- Voynich and Padua — The Northward Route
The Final Idea
A language can tell us where a tradition has been.
It does not always tell us where the page was made.
If Voynich is ever deciphered, the first geographic question should not be “what country speaks this language?” It should be “what route could have brought this linguistic and technical state into the hands of these particular scribes?”