The strangest thing about the title “Voynich Manuscript” is that the manuscript never calls itself that.
Wilfrid Voynich’s name belongs to the twentieth-century history of the object.
Beinecke MS 408 is a modern library shelfmark.
“Herbal section”, “biological section”, “pharmaceutical section” and “recipes” are modern descriptive conveniences.
Even the word “section” can sound more settled than the manuscript deserves.
What did the book call itself?
We do not know.
Who did it say wrote or compiled it?
No securely identified original author statement survives.
Where is the table of contents?
No accepted one has been identified.
Where are the chapter titles?
Some paragraph-final layouts have been called “titles” by researchers, but their function is undeciphered.
Where are the page numbers?
They exist—but they were added later.
Where are the quire numbers?
They exist too—and belong to another organisational layer.
The Voynich Manuscript is therefore heavily structured while withholding much of the explicit navigational apparatus modern readers expect from a book.
This is the paratext problem.
We can see that the manuscript is organised. We cannot yet read the labels by which its original users may have understood that organisation.
Quick Read
One-sentence answer: the Voynich Manuscript has no securely identified original title, named author, conventional table of contents or decoded heading/index system; its modern name and section labels are supplied externally, while later folio and quire numbers, ownership marks and ambiguous paragraph-final “titles” show that several layers of organisation exist without giving us the manuscript’s original self-description.
- “Voynich Manuscript” is a modern name derived from Wilfrid M. Voynich, who acquired it in 1912.
- “Beinecke MS 408” is its modern Yale shelfmark.
- Yale describes the work as written in an unidentified script by an unknown author.
- No original title page has been securely identified.
- No colophon naming an accepted author, date or place has been decoded.
- No accepted table of contents survives in readable form.
- No accepted index system has been decoded.
- Modern section names such as herbal, astronomical, biological/balneological, pharmaceutical and recipes are descriptive visual aliases.
- Those aliases are useful but do not establish original medieval chapter titles or functions.
- Paragraphs are visibly organised and sometimes end in centred, right-justified or special right-side text.
- John Grove called one such layout type “titles”; the label is modern and the function is unknown.
- Quire numbers are later than the principal writing and belong to an organisational/binding layer.
- Folio numbers were added by another later hand.
- Ownership marks and later month names belong to still other layers.
- The absence of readable paratext may reflect loss, unfamiliar conventions, embedded headings, a working/reference genre or simply our inability to recognise the system.
- It does not prove the manuscript was secret, unfinished, meaningless or never meant for readers.
The manuscript may have had navigation.
The problem is that its navigation may be encoded in forms we currently call text, layout or image.
The Modern Name Is Already a Layer of Interpretation
“Voynich Manuscript” feels ancient because it is now famous.
The name is modern.
Wilfrid Voynich acquired the codex in 1912 from the Jesuit collection associated with Frascati.
The manuscript entered modern scholarship through him strongly enough that his surname became the object’s common name.
Yale’s shelfmark, Beinecke MS 408, is even more obviously administrative.
Neither tells us what a fifteenth-century maker called the work.
This is a useful reminder.
A title can become so normal that we forget it is supplied from outside the object.
Before interpreting the manuscript’s internal taxonomy, strip away the modern label mentally.
We are studying an unidentified work now called the Voynich Manuscript.
Yale’s “Unknown Author” Is a Cataloguing Fact, Not a Theory of Authorship
Yale describes the manuscript as written in an unidentified script by an unknown author.
That wording is intentionally conservative.
It does not mean researchers have proven that the author concealed their identity deliberately.
It means no creator has been identified securely enough for the catalogue.
This distinction matters because the manuscript has attracted famous attributions.
Roger Bacon.
John Dee.
Various physicians, alchemists and court figures.
Historical associations can be interesting.
None has become secure authorship.
“Unknown author” preserves exactly the amount of knowledge we currently have.
Author Is Not the Same as Scribe
The Scribe Problem makes the paratext issue even more important.
A scribe writes.
An author composes.
A compiler selects and arranges.
An illustrator draws.
A colourist paints.
One person can perform several roles.
Several people can share one role.
Even if the five-scribe model were confirmed perfectly tomorrow, we still would not know whether there were five authors.
Conversely, one scribe could copy a compilation written by several earlier authors.
The absence of an author statement therefore leaves multiple production architectures open.
handwriting can identify producers of ink; it does not automatically identify producers of ideas.
Where Is the Title Page?
Modern books often put their identity at the front.
Title.
Author.
Publisher.
Date.
Medieval manuscript culture is not organised around the modern printed title page in the same way.
Works can be identified by incipit, rubric, colophon, opening decoration, conventional genre or external catalogue tradition.
Therefore the absence of a modern title page is not surprising by itself.
The more significant fact is that no original readable incipit, rubric or colophon has yet supplied a secure title or author.
The first page contains principal Voynich text, later ownership traces and red painted initials/doodles.
It has not yielded an accepted readable “Here begins the book of…” formula.
That may mean the title is encoded inside Voynichese.
It may mean the original first leaf is missing.
It may mean the work never had the kind of explicit title we expect.
All remain possible.
The Manuscript Could Be Acephalous Without Looking Obviously Broken
An acephalous manuscript is missing its beginning.
If an original title, preface or table of contents occupied lost opening leaves, the present f1r may not be the work’s intended first page.
Do we know that happened?
No.
But current-order uncertainty and known missing material mean the possibility cannot be excluded merely because f1r looks like a plausible page.
This creates a general rule.
Missing paratext should not automatically be interpreted as intentionally absent.
Some may simply be missing.
At the same time, missing leaves cannot be used as a convenient invisible table of contents whenever a theory needs one.
Absence is an open possibility, not recovered content.
Where Are the Chapter Headings?
Modern readers infer large visual sections because illustration families change dramatically.
Plants dominate one region.
Circles and zodiac imagery another.
Human figures and pools another.
Detached plant fragments and vessels another.
Star-marked text another.
This visual segmentation is useful.
But where does the manuscript say “Chapter: Plants”?
We cannot read such a heading.
Indeed, some boundaries modern viewers treat as section changes may be consequences of current binding rather than original conceptual structure.
The modern six-section picture is therefore descriptive.
It is not decoded paratext.
The “Herbal Section” Is a Visual Alias, Not a Medieval Heading
The word “herbal” is one of the most useful and most dangerous names in Voynich studies.
Useful because many pages undeniably depict large plant-like forms beside text.
Dangerous because “herbal” can quietly import function.
A medieval herbal identifies plants and their properties.
Voynich plant pages may do that.
Or they may organise ingredients, symbols, copied visual traditions or another system.
The same applies to “pharmaceutical” and “recipes”.
These labels help humans navigate an unread manuscript.
They should not be turned into evidence that the unread manuscript uses those genre names internally.
A useful paratext article must separate our filing system from theirs.
John Grove’s “Titles” Are Real Layout, Unknown Function
There is one place where the word “title” enters Voynich terminology from inside the page layout.
John Grove used the term for a distinctive paragraph-final configuration.
Ordinary final lines are left-justified and shorter.
Some final lines are centred.
Some are right-justified.
In Grove’s strict “title” layout, the main final line remains left-justified while additional words sit to the right.
Zandbergen’s survey records a small number of pages using this special form.
The layout is real.
The interpretation is not.
It may be a heading.
A tag.
A closing formula.
An identifier.
Or another document role.
“Title” here is a researcher’s nickname for layout, not a translation.
If Grove Titles Are Headings, Their Position Is Unexpected
Modern headings usually precede what they label.
Grove-style material appears at paragraph endings.
That does not rule out a heading function.
Historical manuscripts can place titles or rubrics in diverse locations.
But it creates alternatives.
- label for the paragraph that just ended;
- label for what follows;
- signature-like tag;
- index term;
- summary;
- closing formula.
A real heading model should predict vocabulary and recurrence.
If the right-side words are titles, they should behave more like labels or catalogue entries than ordinary prose.
Their distribution can test the paratext hypothesis without translating them first.
Paragraph Initials May Carry Paratextual Work
Some paragraph initials are enlarged or specially decorated.
This is another place where navigation may be embedded in ordinary script rather than separated as a headline.
An enlarged initial can mark a new entry.
A rubric-like form can signal a new section.
A gallows-rich first token can perform a discourse-opening role.
None is decoded.
But the manuscript clearly uses positional differentiation.
This suggests one possible answer to “where are the headings?”
some of the heading function may be grammatical or graphical rather than written as separate readable titles.
If so, paratext and text are partially fused.
Labels May Be an Indexing System Without Being an Index
Labels occur beside plants, stars, figures, containers and detached plant parts.
They are clearly navigational in one broad sense.
They connect text to visual nodes.
But an index is more specific.
An index normally lets a user retrieve information across the document through repeated keys, ordered entries or references.
Voynich labels are often unique, especially in zodiac contexts.
This makes a conventional repeated-term index less obvious.
They may be names or coordinates.
They may also be local identifiers that participate in cross-reference elsewhere.
No stable manuscript-wide cross-reference syntax has been decoded.
Therefore labels are potential navigation handles.
They are not yet an identified index.
Where Is the Table of Contents?
No accepted table of contents has been recognised in the principal text.
That absence has several possible explanations.
- The book never had one.
- It was written in Voynichese and we do not recognise its function.
- It was on missing leaves.
- The manuscript was assembled from modules and never given one unified contents page.
- Visual sections and quire organisation were enough for intended users.
- Tabs, bookmarks, covers or external finding aids have been lost.
The modular possibility is especially interesting.
If bifolia or packets were produced separately and later gathered, the work may not have begun life as one linear book requiring a contents page.
Again, this is a hypothesis.
The absence of contents cannot prove modular production.
Quire Numbers Are Navigation, but Not Original Content Labels
The manuscript contains quire numbers.
They help reconstruct gatherings and missing material.
They are enormously valuable codicological evidence.
But they belong to an organisational layer later than the principal drawn and written content.
A quire number says:
this gathering occupies this place in an assembled book state.
It does not say:
the original author called this Chapter 8.
Administration and conceptual structure can coincide.
They are not identical.
Folio Numbers Are Even More Clearly Later Navigation
The familiar f1, f2, f3 sequence is essential to modern Voynich discussion.
Those folio numbers were added after the principal content.
They belong to another hand from the quire numbering.
They preserve knowledge of leaves that are now missing.
But they are not the composer’s original page references unless future evidence shows otherwise.
This creates a subtle risk in modern scholarship.
We speak of “f57v” as though that identifier were inherent to the medieval diagram.
It is a later address.
Useful addresses can become cognitively invisible.
The object had another organisational life before our numbering system existed.
Later Month Names Are Paratext About the Images
The readable Romance month names on zodiac pages are a classic example of later paratext.
A later reader saw the images and supplied familiar names.
This is navigation.
It helps identify a diagram.
It also tells us something about reception history.
But it is not original paratext from the principal production event.
The later reader may have been right about the zodiac sign while knowing nothing about the Voynich labels.
This is why the Crib Problem kept later annotations separate from a bilingual key.
Paratext has historical layers just like ink and paint.
The Tepenec Mark Is Ownership Paratext
The faded Jacobus de Tepenec inscription on f1r is another layer.
It tells us about ownership.
It does not tell us the title.
A possession mark can become prominent enough that modern viewers treat it as part of the page’s identity.
Historically, it arrived later.
The same distinction applies to stamps, catalogue marks and other institutional traces.
A codex accumulates metadata through ownership.
Not all metadata comes from the author.
f116v Could Be Paratext Without Being a Colophon
The final verso contains mixed-looking later writing and Voynich-like signs.
Its position at the end creates a strong colophon temptation.
Perhaps it names an owner.
Author.
Place.
Prayer.
Recipe.
Memory note.
But the readings are disputed and the writing layers are not securely aligned as an original concluding statement.
Therefore f116v is important paratextual evidence about later interaction with the book.
It is not an accepted original colophon.
A Working Manual May Need Less Formal Paratext Than a Prestige Book
Genre affects navigation.
A presentation manuscript intended for a patron may carry elaborate title, dedication and heraldry.
A personal working notebook may not.
A technical reference collection can rely on visual categories and memory.
A workshop manual can be meaningful mainly to trained insiders.
A compilation assembled over time may never receive a polished title page.
The Voynich Manuscript’s paratextual sparsity therefore may constrain genre.
But it does not uniquely identify one.
“No visible title” is compatible with several ordinary manuscript contexts.
A Secret Book Does Not Need to Announce That It Is Secret
The missing title also invites secrecy narratives.
No author.
No readable chapter headings.
Unknown script.
Therefore deliberate concealment.
Possibly.
But the same observations can arise from a lost writing tradition, specialised notation, abbreviation, copied material or ordinary paratext loss.
Secrecy requires positive evidence of concealment mechanism.
Paratext absence by itself is weak.
The manuscript is opaque to us.
Opacity and intentional secrecy are not synonyms.
An Unfinished Manuscript Could Also Lack Polished Navigation
Another possibility is incompleteness.
Perhaps headings, rubrication or a contents page were planned but never added.
Colour itself appears to be a later and potentially multi-stage production layer.
Could paratext have been another unfinished layer?
It is possible.
Yet the manuscript also shows extensive finished-looking organisation.
Paragraphs are carefully formed.
Foldouts are engineered.
Large visual systems are complete enough to function as compositions.
Therefore “unfinished” cannot be inferred from missing familiar navigation alone.
A specific unfinished-stage model must identify which planned operation was omitted.
A Modular Compilation May Explain the Weak Global Paratext
The bifolium and production-order articles open another possibility.
Perhaps the work was assembled from packets.
Plant bifolia.
Diagram sheets.
Record sections.
If modules had their own internal organisation, a global table of contents may have been unnecessary until final assembly.
If final assembly was revised, incomplete or later altered, global paratext could be weak while local organisation remained strong.
This hypothesis has a test.
Do individual production packets have stronger internal beginning/ending markers than the transitions between packets?
If yes, modular compilation strengthens.
If one global paratext system crosses every packet consistently, the unified-book model strengthens instead.
The Starred Pages May Have Their Own Embedded Navigation
Quire 20 uses marginal star-like markers to divide short record-like entries.
That is paratextual function even if the star itself carries no linguistic meaning.
The star says, in effect:
a new item begins here.
The entry may not need a written heading because the repeated visual marker and first-token conventions already orient the reader.
This is one reason the manuscript should not be judged only by printed-book navigation standards.
Visual marks can be paratext.
Layout can be paratext.
Character form can be paratext.
Foldouts Have Navigation Built Into Their Mechanics
The Foldouts article established that some information only appears when the book changes physical state.
A fold can hide or reveal a panel.
A large diagram can become one field only when opened fully.
This mechanical access is another paratextual layer.
The physical object tells the reader:
there is more here; unfold me.
We do not know whether the unfolding sequence carried semantics.
But navigation is not limited to written headings.
The book’s mechanics can organise access.
If a Heading System Exists, It Should Form a Distinct Register
Suppose Grove-style titles, centred final lines or special paragraph openings really are headings.
Then their text should behave differently from prose.
Headings tend to be noun-heavy.
Short.
Repeated in navigational contexts.
Less syntactically complete.
A Voynich heading system might show:
- flatter token frequencies;
- label-like initials;
- different q frequency;
- recurrent word families tied to page type;
- special final characters.
These predictions can be tested without knowing the words.
Paratext should have a statistical signature if it is a real document role.
An Index Should Create Cross-Page Retrieval Structure
If the manuscript contains an index hidden in plain sight, it should connect distant content.
An index term repeats or transforms predictably where the indexed object appears.
A coordinate system maps labels to locations.
A cross-reference phrase points from one entry to another.
This creates a powerful criterion.
A proposed “index” token should improve retrieval.
Pages containing the same index key should share an independent visual or textual relation.
If no cross-page relationship emerges, the index interpretation weakens.
Navigation should navigate.
The Missing Paratext Is a Constraint on Decipherments
A proposed translation should eventually identify paratextual roles.
Which words are headings?
Which are labels?
Which start new entries?
Which close them?
Does the first page contain an incipit?
Does the last contain a colophon?
If a decipherment produces fluent paragraphs but never yields any document-management vocabulary anywhere in a long technical codex, that absence deserves explanation.
Conversely, if the proposed headings align exactly with visible special layouts, confidence rises.
A complete solution should recover not only sentences but the book’s interface.
What Survives the Paratext Work
- The common title “Voynich Manuscript” is modern.
- Beinecke MS 408 is a modern shelfmark.
- No original title has been securely identified.
- No accepted original author statement has been decoded.
- No accepted original table of contents or index has been identified.
- Modern section names are descriptive visual aliases rather than recovered medieval headings.
- Special paragraph-final layouts, including Grove-style “titles”, are real document structures with unknown function.
- Paragraph initials and star markers may perform paratextual work.
- Quire numbers, folio numbers, month names and ownership marks belong to later organisational or reception layers.
- Navigation may be partly embedded in layout, labels, geometry and repeated visual markers.
- The absence of familiar paratext does not by itself prove secrecy, incompleteness or non-language.
- A successful decipherment should eventually recover some of the manuscript’s document-management logic.
What Does Not Survive as Established Knowledge
- “Voynich Manuscript” is the original title.
- Beinecke MS 408 is an original manuscript identifier.
- Roger Bacon is the author.
- John Dee is the author or owner by accepted proof.
- The modern six sections are proven original chapters.
- Grove “titles” are proven chapter headings.
- Centred final lines are proven titles.
- f116v is a proven original colophon.
- Later month names are original headings.
- Folio numbers preserve the author’s original sequence.
- The absence of a title proves intentional secrecy.
- Missing opening leaves definitely contained a title page or table of contents.
The organisational layers survive.
The original book title does not.
A Better Paratext Analysis
- Separate modern names from manuscript-native evidence.
- Separate principal writing from later numbering and annotations.
- Map all unusual paragraph-start and paragraph-end layouts.
- Treat Grove “titles” as a hypothesis about function, not a decoded category.
- Compare candidate headings statistically with prose and labels.
- Search for cross-page retrieval relationships before calling anything an index.
- Test whether visual section boundaries align with textual boundary registers.
- Use bifolium and quire structure before trusting current sequence.
- Keep possible missing opening/closing material as uncertainty rather than invented content.
- Require decipherments to recover document roles as well as lexical meanings.
What Would Count as a Real Paratext Breakthrough?
Imagine researchers classify every special paragraph-final and paragraph-initial locus without assigning meaning.
A distinct register emerges.
The register is short, label-like and recurrent at boundaries.
Several recurring forms correlate with known visual page classes.
One form appears before plant pages, another before astronomical diagrams, another at the start of star-marked records.
A developing decipherment then reads those forms as historically plausible rubric or heading vocabulary under the same character values used elsewhere.
Held-out pages are classified correctly from their headings before the illustrations are examined.
That would recover a navigational layer.
Or perhaps the first surviving folio turns out, under a mature decipherment, to contain a conventional incipit and work title whose grammar predicts independent prose.
Then “Voynich Manuscript” would finally acquire an older name.
A paratext breakthrough will not merely tell us what one paragraph says. It will tell us how the original reader knew where they were inside the book.
Primary School: Who Named the Book?
Show a child an untitled notebook.
Write “Anna’s Notebook” on a library card beside it.
Ask whether Anna wrote that title inside the notebook.
Not necessarily.
A useful modern name can come from outside an object.
This is why “Voynich Manuscript” should not be confused with its lost original title.
Lower Secondary: Text, Paratext and Later Annotation
Give students a textbook page.
Ask them to colour-code:
- main text;
- heading;
- page number;
- caption;
- student’s handwritten note.
All are writing.
They do different jobs and belong to different authorship layers.
The exercise models the Voynich distinction among principal script, labels, later foliation and marginalia.
Upper Secondary: Can Layout Be a Heading?
Create a document with no bold titles.
Mark new sections only through spacing, a special opening symbol and a centred final tag.
Students can still navigate once they learn the convention.
This demonstrates how Voynich may contain paratextual functions without modern title formatting.
JC and Adult Readers: Paratext as an Interface Layer
At a higher level, paratext is the interface between content and retrieval.
It maps the document’s internal information to user actions.
- Where do I start?
- What is this section?
- Where does this entry end?
- How do I find it again?
- What does this label point to?
Voynich preserves evidence for several interface operations—paragraph starts, labels, star markers, circles, foldouts—even while the lexical values remain opaque.
A complete model should therefore represent document roles separately from content semantics.
The book may be unreadable to us partly because its interface vocabulary is unreadable too.
A Parent and Teacher Guide
- Ask who supplied every title or section name you use.
- Separate modern catalogue names from original manuscript text.
- Separate author, scribe, illustrator and owner.
- Treat later numbering as navigation history, not original structure.
- Use “herbal”, “pharmaceutical” and “recipes” descriptively rather than as decoded medieval headings.
- Let special layouts be candidate paratext before assigning their meaning.
- Remember that physical folding and visual markers can navigate a document.
- Do not fill missing opening leaves with imagined title pages.
The general lesson is:
the labels we use to understand an unknown object can become so familiar that we accidentally mistake our own filing system for the object’s original self-description.
Reader Checklist: Before You Name a Voynich Section or Heading
- Is the name original or modern?
- Is the text principal or later?
- Is the proposed heading visibly separated from prose?
- Does it occupy a recurring special layout?
- Does its token distribution differ from prose?
- Does it behave like labels elsewhere?
- Does it recur at equivalent document boundaries?
- Does the proposed index term retrieve related pages?
- Is the modern visual section boundary physically original?
- Could missing leaves explain absent paratext?
- Is missing content being invented to support a theory?
- Does a decipherment recover document role under the same lexical rules used for prose?
Frequently Asked Questions
What is the original title of the Voynich Manuscript?
Unknown. “Voynich Manuscript” is a modern name, and Beinecke MS 408 is its modern library shelfmark.
Who wrote it?
No author has been securely identified. The number of principal scribal hands is itself debated and should not be equated automatically with authorship.
Does it have chapter headings?
No heading system has been securely decoded. Special paragraph-final and initial layouts may perform heading-like functions, but their meanings are unknown.
What are Grove “titles”?
A modern descriptive term for a rare paragraph-final layout with additional right-justified words. Their actual manuscript function has not been decoded.
Does Voynich have a table of contents?
No accepted table of contents has been identified in the surviving principal text.
Are the folio numbers original?
No. They are later than the principal manuscript content and belong to a different hand from the quire numbers.
Are the modern sections real?
The visual provinces are real and useful, but names such as herbal, pharmaceutical and recipes are modern descriptive aliases, not decoded original headings.
Could the title have been on a missing page?
Possibly, but there is no evidence that permits us to reconstruct such a page’s content.
What is the strongest current conclusion?
Voynich is richly organised but lacks a securely readable original self-description. Its navigational system may be partly embedded in layout, labels and graphical conventions rather than in familiar title-page and heading structures.
Related eduKateSG Reading
- The Manuscript Before the Mystery
- Marginalia and the Later Hands
- Where Does a Sentence End?
- The Order of Making
- The Crib Problem
Continue Through the Voynich Research Map
Research and Further Reading
- Yale Beinecke — The Beinecke Cipher (Voynich) Manuscript
- Yale — MS 408 Catalogue and Physical Description
- René Zandbergen — Voynich Writing, Labels and Grove Titles
- René Zandbergen — Description and Later Numbering
- René Zandbergen — Origin, Order and Missing Leaves
The Final Idea
We know the Voynich Manuscript by names it did not give itself.
Voynich.
MS 408.
Herbal.
Biological.
Pharmaceutical.
Recipes.
Those names are useful scaffolding around a book whose own interface remains unread.
Yet the manuscript is not disorganised.
Paragraphs begin differently.
Some endings receive special alignment.
Labels attach to visual nodes.
Stars mark entries.
Foldouts reveal larger fields.
Later readers add months, numbers and ownership.
The book has accumulated several navigation systems while hiding the original one from us.
The paratext problem is not that Voynich has no structure. It is that we can see the book telling its reader where things begin, end and belong without yet being able to read the words by which that reader understood the map.