Antidotarium Nicolai as a Voynich Comparator

The Voynich Manuscript has a group of pages that modern readers often call “pharmaceutical” and another group frequently described as “recipes”.

Those names are convenient.

They are not decipherments.

A row of vessel-like drawings does not prove pharmacy. A sequence of short paragraphs marked by stars does not prove recipes. And a plant-like fragment beside a container does not tell us what was put into it.

If we want to test a pharmaceutical interpretation properly, we need a control that is unquestionably pharmaceutical.

The Antidotarium Nicolai gives us one.

The value of the Antidotarium Nicolai for Voynich research is not that the two books look alike. It is that the antidotary shows what information a real medieval compound-medicine system had to carry if a reader was expected to make, recognise, measure and reproduce medicines.

That turns the comparison away from resemblance and toward function.

Quick Read

The Antidotarium Nicolai is a medieval Latin antidotary: a collection centred on compound medicines rather than a book of plant identification alone.
Modern scholarship places it in the Salernitan medical tradition around the late eleventh to early twelfth century, but the identity of the “Nicolaus” associated with it is not secure enough to treat a named historical author as settled fact.
A recent Cambridge history of medieval Mediterranean pharmacology describes “a certain Nicolaus” making a shorter, more user-friendly work by selecting commonly used recipes from the much larger Antidotarium magnum.
The British Library describes the Latin Antidotarium Nicolai as a late-eleventh/early-twelfth-century collection of pharmacopoeial remedies compiled at the School of Salerno.
The work circulated widely and entered later medieval medical education; it was also translated and adapted into vernacular settings.
British Library Harley MS 2374 preserves an anonymous fifteenth-century Middle English translation, demonstrating that the work travelled beyond its original Latin environment.
British Library Egerton MS 747 embeds the Antidotarium Nicolai in a larger medical stack beside a herbal, dosage material, ingredient substitutions, weights and measures, synonyms and an antidotary supplement.
That context is crucial: a compound remedy requires more than ingredient names. It raises problems of quantity, preparation, substitution, terminology and repeatability.
For Voynich research, the antidotary is therefore a control for recipe grammar and executable medical information.
It does not prove that Voynich is an antidotary, that its star paragraphs are recipes, that its vessels are pharmacy jars, or that the Voynich script encodes Latin.

Comparator Firewall

Comparator is not source. Functional analogy is not descent. The Antidotarium Nicolai is used here to define what a known medieval pharmaceutical system had to accomplish—not to assign Voynich a medical identity.

This firewall matters because the comparison is tempting.

Voynich has plant-like material.

It has vessel-like forms.

It has short paragraph units.

Medieval pharmacy used plants, vessels and recipes.

But those shared categories are broad. They generate a hypothesis space, not a conclusion.

First: What Is an Antidotary?

An antidotary is not simply a book about antidotes in the narrow modern sense.

In medieval Latin medical culture, antidotaria could collect recipes for compound medicines made from multiple ingredients. The category sits within a larger pharmacological ecosystem that also includes works on simple drugs, collections of shorter recipes, ingredient glossaries or synonym lists, and texts on weights and measures.

That larger ecosystem is important because no single information form solves every pharmaceutical problem.

A herbal or simple-drug text helps identify and describe ingredients.
An antidotary tells a reader how ingredients enter compound medicines.
A dosage text helps control quantity.
A substitution list handles missing stock.
A synonym list resolves naming variation.
A weights-and-measures text standardises quantities between readers and places.

Egerton MS 747 is especially revealing because several of those layers survive together in one codex.

The Antidotarium Nicolai therefore belongs inside a system rather than floating as an isolated “recipe book”.

The Authorship Problem: Who Was Nicolaus?

Historical writing often turns uncertain names into solid people too quickly.

That is exactly what we should avoid here.

A recent scholarly overview in Drugs in the Medieval Mediterranean refers cautiously to “a certain Nicolaus” who produced a smaller and more usable selection from the much larger Antidotarium magnum.

That wording is valuable.

It separates the historical existence of a work associated with Nicolaus from an overconfident biography of its compiler.

For our purposes, the safest formulation is therefore:

The Antidotarium Nicolai is a Salernitan-associated medieval Latin antidotary traditionally linked to a Nicolaus, whose precise historical identity should not be treated as fully settled.

This is more than editorial caution.

Voynich research has the same problem whenever a marginal name, later annotation or historical owner is promoted into an author or maker without sufficient evidence.

The Date Problem: Use a Range, Not False Precision

The British Library describes the work as a late-eleventh/early-twelfth-century collection associated with the School of Salerno. Modern pharmacological history similarly places the emergence of the Antidotarium magnum and the smaller Antidotarium Nicolai around this transitional period.

That is enough to establish the broad historical order.

It is not a reason to pretend that a precise year of compilation is known.

For comparator work, the range is sufficient: the Antidotarium Nicolai belongs to a pharmacological tradition that had centuries to circulate, be copied, translated, taught and adapted before the fifteenth-century world in which the Voynich parchment was produced.

Compression: From a Huge Recipe Universe to a Usable Working Set

The Cambridge overview gives us one of the most interesting facts about the work.

The earlier Antidotarium magnum could be enormous, with some manuscript versions containing very large numbers of compound recipes. The smaller Antidotarium Nicolai was created by selecting recipes in common use.

This is information compression.

Not lossy compression in the careless sense.

Operational compression.

A large pharmacological universe was reduced into a smaller working repertoire that could be copied, learned, consulted and used more efficiently.

This matters for manuscript architecture.

A reference book intended to hold everything can tolerate breadth.

A practical educational or professional book benefits from selection.

The difference can change vocabulary, repetition, ordering and page density even when both books belong to the same domain.

That gives Voynich research another warning: text statistics may reflect the role of a book within a knowledge system, not merely its language.

A Compound Remedy Is an Executable Object

A narrative can remain meaningful even if some details are vague.

A compound medicine is less forgiving.

To reproduce a preparation, a reader may need to resolve several variables:

which ingredients;
which form of each ingredient;
how much of each;
which sequence of operations;
whether ingredients are ground, boiled, dissolved, strained, mixed or otherwise processed;
what carrier or medium is used;
how the final preparation is administered;
for what indication or context.

Not every surviving medieval recipe states every variable with modern laboratory precision. Manuscripts assume trained readers, inherited conventions and tacit knowledge.

But the deeper requirement remains:

If a text is meant to produce a repeatable compound, it must encode enough structure for its intended expert reader to constrain the action.

That is the operational standard a Voynich “recipe” hypothesis should face.

Recipe Grammar Should Be More Constrained Than Ordinary Prose

Recipes have recurring jobs.

Name the preparation.

List or introduce ingredients.

Specify quantities or proportions.

Encode actions.

Close with use, storage or administration.

These recurring functions can create detectable structural regularity even before every word is understood.

In a readable corpus, we can ask whether certain words cluster in ingredient positions, whether quantity markers recur, whether imperative or procedural forms favour particular positions, and whether recipe openings and closings have stable templates.

That gives us a test for Voynich star paragraphs.

If they really are recipes, they should not merely look like short independent units.

They should reveal a recurring internal grammar compatible with executable procedures.

The Quantity Channel Is a Hard Requirement

The Voynich Manuscript does not present us with a securely recognised ordinary numeral system in its main script.

That is not proof against pharmacy.

Quantities can be expressed by words, abbreviations, specialised signs, repeated units or conventions that an outsider does not recognise.

But a pharmaceutical interpretation has to confront the problem.

Egerton MS 747 is powerful here because the Antidotarium Nicolai in that manuscript is followed by material on dosage, ingredient substitution and weights and measures.

The codex itself tells us that quantitative and equivalence problems belonged near compound-drug practice.

A Voynich recipe theory that never identifies a plausible quantity mechanism remains operationally incomplete.

Standardisation Is About Reproducibility, Not Modern Regulation

It is tempting to call an influential medieval antidotary a “pharmacopoeia” in the fully modern regulatory sense.

That can overstate the historical situation.

The safer and more useful point is that a widely circulated antidotary can create practical standardisation.

If many learners and practitioners encounter the same named preparations, ingredient sets and proportions, the text becomes a shared reference layer.

It can reduce variation between:

teachers and students;
physicians and apothecaries;
one manuscript copy and another;
one city and another;
one generation of practice and the next.

That does not make medieval pharmacy uniform.

It creates a common platform against which local variation can occur.

Education Changes the Text’s Function

The British Library notes that the work became an important part of later medieval university medical curricula.

That matters because a teaching text has different selection pressures from a private notebook.

A curriculum benefits from:

a manageable repertoire;
stable names;
repeatable ordering;
shared formulae;
material that can be examined, discussed and copied.

Therefore frequency and repetition in a manuscript can be driven partly by pedagogical function.

This is another control Voynich analysis should keep available. Repetitive structure does not uniquely identify cipher, language or generation. It may reflect a constrained technical or instructional genre.

Harley MS 2374: The Work Crosses Into Middle English

British Library Harley MS 2374 preserves an anonymous fifteenth-century Middle English translation of the Latin Antidotarium Nicolai.

This is important for two reasons.

First, it demonstrates transmission.

The pharmaceutical knowledge was not locked permanently to one Latin wording.

Second, it separates knowledge identity from surface language.

The same broad therapeutic tradition can survive translation while vocabulary, syntax and spelling change dramatically.

That is a critical comparator principle.

If two manuscripts perform the same informational job, they do not need to look linguistically identical. Conversely, linguistic or visual resemblance alone does not prove that they perform the same job.

Translation Is Also Transformation

When a technical work moves from Latin into a vernacular language, several things can happen.

ingredient names acquire vernacular equivalents;
Latin technical terms are retained where no stable local equivalent exists;
weights may be glossed or adapted;
instructions may become more explicit for a new audience;
local ingredients or practices may enter marginally or through later additions;
copyists can reorganise or abridge material.

This means a genuine historical transmission chain can contain substantial textual drift while preserving functional continuity.

Voynich source theories therefore need something stronger than exact word matching.

They need a mechanism of transformation.

The Antidotarium Inside Egerton 747 Is More Informative Than the Antidotarium Alone

One of the strongest pieces of context comes from British Library Egerton MS 747.

In that codex the Antidotarium Nicolai is physically adjacent to other pharmacological problem-solvers:

De dosibus medicinarum;
a quid pro quo list for substituting unavailable ingredients;
fragments on weights and measures;
a long synonym list for plants and medical ingredients;
a supplement to the antidotary.

This is extraordinary comparator material because the physical manuscript reveals the adjacency of real operational needs.

Compound formulae create questions.

How much?

What if I cannot obtain this ingredient?

What does this regional name correspond to?

Which unit does this writer mean?

Egerton stores answers near the recipes.

The Supply Chain Is Hidden Inside the Recipe

A medieval compound medicine could combine local and imported substances.

Recent scholarship on Mediterranean pharmacology emphasises the movement of drugs and knowledge across Byzantine, Islamicate, Jewish and Latin worlds, as well as the movement of substances from much farther east.

That means every ingredient list silently depends on:

trade;
authentication;
storage;
substitution;
linguistic equivalence;
availability;
price;
quality.

The recipe is therefore only the visible tip of a larger pharmaceutical system.

This is why apparently peripheral lists of synonyms or substitutes are not peripheral at all.

They keep the prescription executable when the world changes underneath it.

One Historical Recipe Also Reminds Us Not to Romanticise the Past

Modern pharmacological research has examined one well-known sedative preparation transmitted in the Antidotarium Nicolai, sometimes translated as the “Great Rest”. The historical formula involved potent ingredients including opium, henbane and mandrake.

The point here is not to reproduce or recommend the preparation.

It is the opposite.

Historical medical texts are evidence about past knowledge systems, not instructions for modern self-treatment.

Pre-modern remedies can include pharmacologically active, toxic or poorly standardised materials. Comparator research should study their structure and history without turning manuscript medicine into contemporary medical advice.

What a Real Recipe Signal Should Look Like in Voynich

If the Voynich star paragraphs encode recipes, we should be able to ask for structural predictions before reading their meaning.

Do paragraph openings contain a restricted set of formulaic elements?
Do some token classes recur in positions compatible with ingredient introduction?
Is there a plausible quantity channel?
Do action-like token classes occur in predictable positions?
Do related preparations share repeated frames with variable ingredient slots?
Do closing positions have a different token distribution from middles?
Are paragraph lengths related to the number of illustrated components elsewhere?
Can the same grammar be detected on pages not used to build the model?

These questions can be tested without deciding whether one Voynich word means “grind” or another means “honey”.

That is exactly why syntax-before-semantics matters.

A Vessel Is Not a Recipe

Voynich’s so-called pharmaceutical pages contain container-like objects and plant fragments.

Containers certainly belong to pharmacy.

They also belong to kitchens, workshops, trade, storage, perfumery, alchemy, cosmetics and ordinary domestic life.

Therefore container shape is low-specificity evidence unless other systems cohere with it.

A pharmaceutical interpretation strengthens if containers correlate with:

distinct ingredient-like labels;
recurrent recipe grammar;
quantity markers;
stable preparation categories;
cross-references to plant or substance pages.

One icon category is not enough.

A Star Is Not a Bullet Point Until We Prove Its Function

Voynich’s final text-heavy pages frequently begin short units with star-like marks.

Modern eyes naturally treat them as bullets.

They may be.

They may also be paragraph markers, indexing devices, decorative separators, category symbols, astronomical signs or something else.

The Antidotarium Nicolai gives us a useful discipline: if the units are recipes, the internal text should behave like recipes regardless of what the stars look like.

Form should be tested against function.

Could Recipe Structure Explain Voynich Predictability?

Technical genres can be predictable.

A recipe corpus repeats ingredient categories, quantities, preparation verbs and formulaic transitions.

This can lower local uncertainty compared with unconstrained narrative prose.

That makes recipe language a legitimate control for Voynich entropy work.

But the mechanism must reproduce more than low entropy.

word-length structure;
line effects;
Currier differences;
word families;
labels versus running text;
section-specific vocabulary;
paragraph architecture.

If a recipe corpus matches one statistic while failing the rest, the comparison has done its job: it has constrained the theory rather than confirmed it.

The Strongest Medical Hypothesis Would Join Several Voynich Sections

A serious pharmaceutical model would not explain the “recipe” pages in isolation.

It would try to explain a workflow.

substance identification → preparation or classification → compounding → quantity → administration → retrieval

If different Voynich page types participate in one domain, some cross-section relationships should emerge.

Words or labels associated with a plant-like entity might recur in a preparation context.

Recipe-like units might systematically draw from vocabulary concentrated elsewhere.

Quantity-like tokens might cluster in pharmaceutical and recipe contexts.

A proposed medical theory gains strength when separate parts of the manuscript constrain each other.

What the Antidotarium Nicolai Can Legitimately Teach Voynich Research

A medieval pharmaceutical text can be deliberately compressed into a smaller working repertoire.
Compound-remedy writing has stronger operational constraints than unconstrained prose.
A usable recipe system must manage ingredient identity, quantity and procedure somehow.
Standard references can reduce variation without eliminating local practice.
Teaching and professional use can generate repeated formulae and stable textual structures.
Technical knowledge can move across languages while retaining functional identity.
An antidotary can sit inside a larger stack containing dosage, substitutions, weights, measures and synonyms.
Supply-chain constraints are part of pharmaceutical information architecture.
A medical interpretation becomes stronger when it predicts cross-section relationships rather than explaining one visual motif.

What the Antidotarium Nicolai Does Not Prove

that Voynich is an antidotary;
that Voynich was made at Salerno;
that its star paragraphs are recipes;
that its vessel-like drawings are apothecary containers;
that its script represents Latin;
that its word families are ingredient names;
that any Voynich section is a translation of the Antidotarium Nicolai;
that the historical “Nicolaus” can be identified with certainty;
that a medieval remedy should be reproduced or used as modern medical treatment.

A Pre-Registered Voynich Recipe Test

Define the candidate recipe corpus before inspecting comparators.
Freeze paragraph boundaries and transcription rules.
Learn positional token classes without assigning meanings.
Search for repeated internal frames with variable slots.
Identify candidate quantity channels independently from ingredient hypotheses.
Compare against readable medieval antidotaria, receptaria and non-recipe prose controls.
Score missing expected structure as negative evidence.
Test the learned grammar on held-out Voynich pages.
Only after structural validation should semantic guesses be introduced.

This sequence protects the analysis from one of Voynich research’s oldest traps: deciding what the text says first and discovering “evidence” for it afterward.

Primary School: Why Instructions Need Enough Information

Ask students to write instructions for making a cup of chocolate milk without using quantities.

Compare the results.

Then add measured quantities and a sequence.

The exercise shows why an executable instruction is different from a description.

Lower Secondary: Find the Recipe Grammar Before the Meaning

Give students several unfamiliar but readable recipes with ingredient names replaced by symbols.

Can they still identify quantity positions, action positions and repeated openings?

This models the difference between structural inference and translation.

Upper Secondary: Compare Genres With the Same Statistics

Compare recipe text, narrative prose and generated template text after matching paragraph length.

Measure repeated words, positional regularity and vocabulary growth.

Students learn why one statistic rarely identifies a mechanism by itself.

JC and Adult Readers: Recipe Text as a Typed Program

At a higher level, a compound recipe can be represented as a typed procedure.

Ingredients are entities.

Weights are quantities.

Preparation verbs are transformations.

Containers and media are state environments.

Administration is an action boundary.

A text that truly encodes recipes should therefore contain information sufficient to constrain transitions between those types.

This does not mean a medieval recipe behaves like modern software.

It means reproducible action leaves structural demands in language.

Research and Primary Sources

Continue Through the Voynich Comparator Library

The Final Idea

The Antidotarium Nicolai gives Voynich research a more demanding question.

Not:

Do those Voynich pages look like pharmacy?

But:

If those pages encode compound medicines, where is the information architecture that makes compound medicine executable?

Where are the ingredient roles?

Where is quantity?

Where are transformations?

Where are the repeated frames?

Where is the relationship between a substance shown in one part of the book and a preparation used in another?

A real pharmaceutical interpretation should make those questions easier to answer, not merely rename the mystery.

The Antidotarium Nicolai is useful not because it solves Voynich, but because it tells us what a solution claiming “pharmacy” should eventually have to explain.

Voynich Comparator | Antidotarium Nicolai: How Medieval Pharmacy Turned Compound Remedies Into a Repeatable System