Voynich and Padua | Francesco Zago’s Library Inventory of 9 May 1404

Sometimes history gives us a date.

Sometimes it gives us an object.

And occasionally it gives us something even more useful: a dated snapshot of a system in operation.

On 9 May 1404, Francesco Zago recorded a list of books belonging to Francesco II “il Novello” da Carrara, the last lord of Padua.

For Voynich research, this document is important for a reason that is easy to misunderstand.

Zago’s inventory does not contain a demonstrated Voynich Manuscript. Its value is that it turns the late Carrara intellectual world from atmosphere into partially enumerated evidence.

We can ask what books were present, how they were described, what subjects dominated, which books moved between readers, which surviving manuscripts can be identified, what vanished, and what kind of documentary description a genuinely relevant unknown codex would have needed to leave.

This is where the Padua hypothesis becomes archival rather than aesthetic.

Quick Read

Francesco Zago compiled the core list on 9 May 1404.
Sarah R. Kyle describes the initial record as containing 57 manuscripts.
Four additional volumes were entered in succeeding months, producing the 61-volume total often cited in later summaries.
Zago was an official deputy of the Carrara massaria, the office involved in managing Padua’s revenues.
Entries could record author, title or brief contents, language, binding, format and identifying marks.
The reverse of the record preserved loan information from the period before Venetian seizure of Carrara effects and property.
Kyle reports that more than two-thirds of the listed titles concerned the theory or practice of medicine, especially traditions with Arabic medical ancestry.
The Carrara Herbal can be identified in the list through the entry commonly rendered as Serapiom in volgare.
The whereabouts of most books on the list remain unknown.
No entry has been demonstrated to describe Beinecke MS 408.

Why an Inventory Is Different From a Library

The first trap is to call Zago’s list “the Carrara library” and quietly assume that the two are identical.

They are not.

An inventory is a representation of a collection under particular administrative conditions. It records what the person making the document had reason and opportunity to record.

A book can exist outside the inventory because it is:

on loan;
stored elsewhere;
personally held by another family member;
newly acquired;
not considered part of the relevant administrative category;
missed;
described under a title modern scholars do not recognise;
present as loose material rather than as the sort of codex being inventoried.

Conversely, an inventory entry can survive when the actual book does not.

So Zago gives us neither perfect presence nor perfect absence.

He gives us a bounded historical observation.

The Document Has a Precise Administrative Setting

Kyle identifies Zago as the official deputy of the Carrara massaria, an office managing revenues. The inventory formed part of a consegna, a note of consignment or delivery.

This administrative origin matters.

The record was not created six centuries later by a bibliophile trying to reconstruct Francesco Novello’s reading tastes. It belonged to the machinery that managed possessions.

That makes apparently mundane details unusually powerful.

Bindings matter.

Format matters.

Marks matter.

Loan notes matter.

The inventory sees books as trackable objects, not merely as texts.

57 Books, Then Four More

The number requires care because two totals circulate.

Kyle’s dissertation states that Zago’s list of 9 May 1404 contains 57 manuscripts. Four further volumes were entered in the succeeding months, brought forward by priests identified as Cristoforo and Brussano.

That produces the later total of 61 found in summaries such as Treccani’s biography of Francesco Novello.

There is no contradiction if we preserve the event sequence:

9 May state = 57 recorded manuscripts → later additions = 4 → expanded record = 61.

This small distinction is a good example of why Voynich research needs temporal precision. A source can be correct at one moment and another source correct about the evolved record.

What Zago Recorded

Kyle describes the entries as capable of including:

the author’s name;
the title;
a brief description of contents;
the language of the text;
a description of the binding;
the manuscript’s format;
identifying markings such as coloured symbols.

For a historian, that is much richer than a list of titles.

A title identifies intellectual content.

A binding identifies an object state.

A language identifies a representation choice.

A mark can identify an individual volume even when several books share similar contents.

In modern terms, Zago was recording multiple metadata fields.

That is why the document can support something closer to entity resolution than simple title matching.

The Most Striking Fact: Medicine Dominates

The collection is remarkable because Francesco Novello was not a university-trained physician.

Yet Kyle reports that over two-thirds of the works on Zago’s list concern health and medicine, especially medical traditions with substantial Arabic ancestry.

This is not a minor thematic preference.

It makes medicine one of the strongest organising features of the surviving snapshot of the prince’s library.

Treccani highlights works associated with important Paduan physicians, including the Consilia of Marsilio Santasofia and material associated with Pietro da Pernumia.

Kyle’s study places the Carrara Herbal inside this much larger intellectual project, not as an isolated curiosity.

That changes how we should think about the Paduan court.

The court did not merely possess one beautiful herbal. It sat beside a university and maintained a collection in which medical knowledge occupied an extraordinary proportion of the recorded shelf-space.

The Arabic Medical Layer Matters — But Carefully

The Carrara Herbal itself is an excellent example of long-range knowledge transmission.

The underlying materia medica tradition is associated with Serapion the Younger. It travelled through translation into Latin and then into a Paduan vernacular form.

That route is historically important because it shows why geographical labels can mislead.

A Paduan book can contain Arabic-derived medicine.

A Latin source can become a vernacular interface.

A courtly manuscript can encode university-relevant knowledge.

Therefore, if Voynich eventually proves connected to a Paduan intellectual environment, we should not expect its knowledge necessarily to have originated in Padua.

The Carrara Herbal Is the Inventory’s Great Reality Check

Most inventory research suffers from a basic problem: descriptions survive, objects disappear.

Egerton MS 2020 lets us test what successful identification looks like.

Kyle identifies the inventory entry Serapiom in volgare with the manuscript now called the Carrara Herbal. The surviving codex independently carries Carrara heraldry and a Paduan scribal colophon, and the British Library locates its production in Padua around 1390–1404.

That is the standard we want.

Not:

the title sounds mysterious, so perhaps it is Voynich.

But:

documentary description + surviving object features + independent ownership/provenance evidence + compatible date and content.

A genuine Voynich match should need comparable convergence.

Books Were Moving Inside the Collection

One of the most useful details in Kyle’s reconstruction is that the reverse of the inventory preserves a history of loans.

For example, Kyle notes that Milone, a natural son of Francesco Novello, borrowed a Tesaurus pauperum on 4 August and returned it on 15 August.

The exact book is less important here than the behaviour.

The library had circulation.

Books crossed hands.

A volume’s location on one day was not necessarily its location on another.

This has two consequences for Voynich research.

An inventory should not be treated as a static map of every book’s physical position.
Knowledge transmission can occur through temporary access without ownership transfer.

A student, courtier, physician or family member does not have to own a book to copy from it.

1404 Is a State Snapshot Just Before Catastrophe

Zago’s date becomes more significant because of what follows.

The Carrara–Venetian war intensifies. Padua falls in 1405. Francesco Novello and his sons die in Venetian captivity in January 1406.

Kyle describes the loan record as preceding the seizure of Carrara effects and property by Venice’s Council of Ten in 1406.

That means the inventory sits at a rare boundary:

we see part of the collection while its political owner still rules, shortly before that ownership system is destroyed.

For provenance research, this is exactly the kind of historical discontinuity that can create dispersal.

But again, dispersal is a mechanism, not a licence to insert any later mysterious manuscript into the gap.

Most of the Books Are Still Missing

Kyle notes that the whereabouts of most books listed by Zago remain unknown.

This is both exciting and dangerous.

It is exciting because the collection is not historically exhausted. There are unresolved identities.

It is dangerous because every unidentified entry becomes a temptation.

A missing book is not evidence for whichever missing book we hope to find.

Unknown does not mean Voynich.

Unknown means unknown.

Does the Voynich Manuscript Appear in Zago’s List?

At present, the responsible answer is:

No entry has been demonstrated to be Beinecke MS 408.

This sentence should remain fixed until evidence changes it.

There are several logically possible reasons:

Voynich was not in Francesco Novello’s collection.
The parchment had not yet been made into the completed manuscript.
The relevant object was outside the inventoried subset.
A record could describe an object in terms that do not allow modern identification.
The intellectual ancestry could be Paduan while the physical manuscript never belonged to the Carrara.

All are possible.

None may be selected merely because it protects Padua.

A Better Inventory-Matching Protocol

If we search Zago’s inventory for something relevant to Voynich, a useful match should be evaluated across several fields.

1. Contents

Does the description imply a text family compatible with what can actually be established from the candidate manuscript?

2. Language

If an entry explicitly says Latin or vernacular Italian, an unidentified script creates an immediate problem unless the inventory description refers to underlying content rather than visible script.

3. Format

Can manuscript size or format plausibly match? If original dimensions or binding state have changed, can that change be independently demonstrated?

4. Binding

This field is difficult for Voynich because its current binding is later. A binding mismatch therefore cannot be hand-waved away, but neither can the present binding be treated automatically as original.

5. Identifying Marks

A distinctive mark would be powerful if it survived. Absence is harder to interpret if leaves, covers or endpapers were lost.

6. Ownership Path

A candidate match becomes serious only if we can explain how the object moves from the Carrara context toward a later documented point.

7. Negative Evidence

What should be present if the match is true but is not present?

This last question prevents a theory from accumulating only supportive details.

Why a Generic “Secret Book” Is Not Enough

Historical catalogues frequently contain abbreviated, vague or unfamiliar descriptions.

That can make an untitled or poorly identified item feel tantalising.

But the more generic the entry, the more objects it could match.

A phrase such as “a book of herbs” has low discriminating power.

An entry specifying author, language, size, binding and a particular identifying mark has much higher discriminating power.

Mystery is not specificity.

For that reason, this series will not use vague inventory wording as provenance proof.

What the Inventory Really Adds to the Padua Hypothesis

Even without a Voynich match, Zago changes the landscape in five major ways.

It dates the collection. We are not reconstructing an abstract “late medieval” library.
It demonstrates medical density. Medicine is not inferred from one surviving herbal.
It demonstrates object management. Books are described by physical and textual metadata.
It demonstrates circulation. Loan records show the collection was used and moved.
It creates a dispersal boundary. The record precedes the destruction of Carrara rule and Venetian seizure of property.

These are independent of Voynich.

That is precisely why they are useful.

The Carrara Library Was Rebuilt After an Earlier Dispersal

Treccani adds another important layer.

When Padua fell to the Visconti in 1388, valuable Carrara manuscripts were taken toward Pavia. After Francesco Novello recovered Padua, he worked to rebuild the library by recovering some volumes and acquiring others.

Therefore, by 1404 the collection Zago records is already a post-disruption collection.

This means the Carrara library itself had a history of loss, recovery and recomposition before the final 1405 collapse.

That is an excellent reminder that manuscript collections should be modelled as changing states rather than permanent shelves.

A Library Can Preserve a Knowledge Cloud Without Preserving Every Book

Suppose forty medical books disappear.

The knowledge network does not necessarily disappear with them.

Teachers may have read them.

Students may have copied from them.

Physicians may have internalised methods.

Recipes may have entered practical notebooks.

Artists may have reused visual models.

Another copy of the same text may survive elsewhere.

This makes the inventory useful even where individual object identity fails.

It helps reconstruct the available intellectual field.

What We Should Search Next

The inventory produces concrete research jobs.

Build a normalized table of all 57 original entries and the four later additions.
Separate securely identified surviving books from probable identifications and unresolved entries.
Record the exact language, binding, format and identifying-mark metadata for every entry.
Map all named authors and works to the University of Padua, court, physicians and wider transmission networks.
Trace surviving copies of the same works elsewhere to establish how their diagrams, abbreviations and page architectures vary.
Reconstruct the loan network from the reverse of the record.
Trace known Venetian seizures and subsequent ownership where records survive.
Compare every tempting Voynich candidate description against non-Voynich manuscripts that could match equally well.
Keep unresolved entries unresolved until independent evidence narrows them.

The Voynich Test

If someone proposes that a particular Zago entry is Voynich, ask four questions immediately.

Why this entry rather than every other medieval book that could fit the wording?
Which independent physical feature of Beinecke MS 408 matches the inventory description?
What historical route connects the Carrara possession to the later manuscript?
What observation would falsify the identification?

If those questions have no answers, the entry is a lead, not a provenance claim.

Research Sources

Continue the Series

The Final Idea

Francesco Zago does not hand us the answer to Voynich.

He hands us something more disciplined.

A date.

A count.

A metadata system.

A medically concentrated collection.

A record of books moving between readers.

A surviving medical manuscript that can actually be matched back to the list.

And a large unresolved remainder.

The correct question is not “Which mysterious Zago entry can we call Voynich?” It is “What evidential standard lets us identify one surviving object from a 1404 administrative record, and can Beinecke MS 408 ever meet that standard?”