Voynich Comparator | British Library Add MS 8928: A Tenth-Century Medical Stack Where the Herbal Is Only One Layer

The word herbal can make a medieval medical manuscript look simpler than it was.

We imagine a book of plants.

One plant.

One picture.

One remedy.

Then we meet British Library Add MS 8928.

The manuscript was produced in the tenth century, most likely in southern Italy, with decorative features that may point toward Capua. Its longest herbal component is the Herbarius attributed to Apuleius Platonicus, accompanied by small plant drawings.

But that is only one layer.

The codex also contains Isidore, material attributed to Galen and Hippocrates, gynaecological writing, diet, recipes, animal medicine, Pseudo-Dioscorides and other short medical texts. Recipes appear in contemporary and slightly later hands.

This is not a tidy modern textbook.

It is a medical stack.

Add MS 8928 is valuable to Voynich research because it shows, centuries earlier, that plant knowledge, animal medicine, prognosis, diet, recipes and inherited authorities could coexist inside one working manuscript environment.

Quick Read

British Library Add MS 8928 dates to the tenth century.
It was probably produced in southern Italy; decorative evidence has been associated with a Beneventan/Capuan environment.
It is written in Latin on parchment.
It includes an illustrated Apuleius Platonicus Herbarius.
Plant drawings accompany the herbal material.
The codex also includes Isidore of Seville, texts attributed to Hippocrates and Galen, dietary material, gynaecology, animal medicines, Pseudo-Dioscorides and recipes.
Recipes were added by contemporary and slightly later hands.
The manuscript therefore behaves like a layered medical working environment rather than a single-purpose plant atlas.
Its great chronological distance from Voynich makes it useful as a deep historical control, not a source candidate.
It demonstrates how long-lived medical traditions can be recompiled and re-hosted repeatedly.

Comparator Firewall

Tenth-century southern Italy is not a Voynich provenance proposal. Add MS 8928 is a structural control for what a layered medical manuscript can contain.

The chronological distance is a feature here.

If a manuscript behaviour exists centuries before Voynich, we should not treat that behaviour as a unique fifteenth-century clue.

The Herbal Is Embedded in a Medical Ecosystem

The British Library catalogue does not describe Add MS 8928 as one continuous herbal.

It describes a compendium.

That distinction matters.

The Apuleius herbal sits beside short medical treatises, prognosis, dietetics, bodily knowledge and animal materia medica.

So if we removed the readable Latin and looked only at the page types, a modern viewer might wrongly classify different regions as separate subjects.

The historical reader could experience them as one medical resource.

This Is a Deep Control for the Voynich “Six Sections” Problem

The Voynich is commonly divided into herbal, astronomical/astrological, biological, cosmological, pharmaceutical and recipe-like sections.

That partition is useful descriptively.

But Add MS 8928 gives us a warning from centuries earlier.

A medical codex can combine several bodies of knowledge whose modern subject labels look separate.

Therefore visual heterogeneity is not sufficient to infer multiple unrelated projects.

Nor does a historical compendium prove Voynich unity.

It tells us that integration is a hypothesis to test, not a surprise to reject.

Pseudo-Apuleius Is an Example of Extreme Textual Longevity

The herbal tradition attributed to Apuleius circulated for centuries in multiple manuscript families, languages and visual forms.

Add MS 8928 is one witness inside that much longer history.

This matters for Voynich comparison because an image or textual structure seen in a fifteenth-century book need not have originated in the fifteenth century.

Medieval manuscripts routinely carry old knowledge in new physical hosts.

The date of the manuscript and the age of the tradition are separate variables.

A fifteenth-century page can carry a much older information lineage. Parchment date is not information-age date.

The Plant Image Does Not Live Alone

The Apuleius section includes plant drawings, but the codex around it also contains animal medicine and other medical material.

This is important for the Voynich pharmaceutical and herbal relationship.

Plant representation can be one entry point into a larger materia-medica system rather than the whole purpose of the book.

A root, leaf or flower may be an identifier.

The operational knowledge may live elsewhere—in recipes, dosage, substitution, compound preparation or disease-specific instruction.

This is exactly why image-only comparison cannot recover book function.

Recipes in Additional Hands Show the Book Stayed Open to Use

The British Library notes recipes written in contemporary and slightly later hands throughout the manuscript.

That transforms the codex from a frozen copy into a living object.

A reader could add material.

A medical manuscript could accumulate useful instructions over time.

This gives us a comparator for the Voynich marginalia and later-hand problem.

Not every added note is contamination.

In working manuscripts, addition can be part of use history.

Authority Is Layered Rather Than Singular

Add MS 8928 brings together materials associated with several major names and traditions.

Isidore.

Hippocrates.

Galen.

Apuleius.

Dioscorides.

Sextus Placitus.

Some attributions are pseudonymous or transmitted through complex traditions.

The important structural lesson is that medieval medical authority can be assembled.

A compiler does not need one source author.

The book can be a convergence point.

This Weakens the Search for a Single Voynich Exemplar

“Which book was Voynich copied from?” may be the wrong granularity.

Add MS 8928 makes another architecture historically ordinary:

multiple inherited source traditions → one new manuscript host → later additions

If Voynich is compilatory, different sections or even different layers on one page may have different source ancestries.

That remains a hypothesis.

The comparator tells us the architecture is historically possible.

The Oldest Layer May Not Look Old

Knowledge can be repeatedly rewritten into contemporary scripts and page conventions.

Therefore archaic content does not require archaic handwriting.

This is another reason visual palaeography of the Voynich script cannot directly date the age of any underlying textual tradition.

A fifteenth-century scribe could encode or copy material whose intellectual ancestry was centuries older.

Southern Italy Is a Useful Geographic Negative Control

The current Padua hypothesis naturally increases attention to northern Italy.

That creates a cognitive risk.

The comparator library starts filling with northern Italian books because that is where we are looking.

Add MS 8928 deliberately pushes outside that field.

It reminds us that complex medical compilation and illustrated herbal practice were not Paduan inventions.

If a feature also appears in a tenth-century southern Italian manuscript, it should receive less localisation weight in a Padua argument.

What Add MS 8928 Can Legitimately Teach Voynich Research

A medicinal herbal can be one module inside a broader medical compendium.
Plant, animal, dietetic, prognostic and recipe knowledge can coexist.
Ancient and late-antique traditions can survive in much later manuscript hosts.
Later users can add recipes to an inherited book.
One codex can converge material from multiple authorities.
Manuscript date and information-lineage date must be separated.
Mixed medical content is not uniquely northern Italian or fifteenth-century.
Searching for one complete Voynich exemplar may be too restrictive.

What It Does Not Prove

that Voynich descends from Pseudo-Apuleius;
that Voynich was made in southern Italy;
that every Voynich plant has an ancient herbal ancestor;
that all Voynich sections belong to one medical compendium;
that later additions explain Voynich Currier variation;
that a broad medical stack identifies the Voynich’s purpose.

A Deep-Time Comparator Test

Build image/text feature vectors across dated Pseudo-Apuleius witnesses.
Measure what survives across centuries of copying.
Separate iconographic inheritance from local artistic style.
Track when new recipes and marginal material enter manuscripts.
Measure whether ordering is more stable than image form.
Compare those survival rates with proposed Voynich correspondences.
Downweight features common across many centuries and regions.
Upweight rare coupled features that remain genealogically informative.

Student Lens: A Book Can Be a Stack

Primary: imagine a notebook containing plant facts, animal facts, recipes and health tips. Different page types can still serve one larger purpose.

Lower Secondary: distinguish a book’s physical date from the age of the ideas copied into it.

Upper Secondary: map source traditions as modules converging into a new compilation.

JC and adult readers: treat textual ancestry as a directed acyclic graph rather than forcing every manuscript into one parent-child chain.

Research Sources

Continue the Comparator Route

The Final Idea

Add MS 8928 is nearly half a millennium older than the Voynich parchment.

That distance is precisely why it matters.

It removes features from the category of “special fifteenth-century clue.”

Plant images inside medicine are old.

Mixed medical compilations are old.

Multiple authorities inside one codex are old.

Readers adding recipes are old.

So the Voynich problem gets narrower.

We stop treating inherited medieval book behaviours as mysteries.

We reserve the mystery for what remains unexplained after those behaviours are controlled.

A good deep-time comparator does not tell us where Voynich came from. It tells us which parts of Voynich do not need a new explanation because medieval medicine had already been doing them for centuries.