The Voynich Manuscript is usually introduced as if somebody already found its table of contents.
Herbal.
Astronomical and astrological.
Balneological.
Cosmological.
Pharmaceutical.
Recipes.
Six sections. Six labels. Six apparent subjects.
It is a wonderfully convenient way to enter an inconvenient manuscript.
It is also one of the places where modern description can quietly harden into medieval certainty.
The six sections are real as a modern navigational system. They are not six recovered chapter titles.
That distinction is the entire reason for this article.
Quick Read
- Yale’s public description usefully presents six broad illustration groups: plants; astronomical/astrological drawings; baths and bathing; cosmological medallions; pharmaceutical herbs/roots; and mostly unillustrated text possibly representing recipes.
- Those groups are based mainly on what the pages look like, because the writing remains unread.
- The labels are therefore descriptive conventions, not decoded semantic chapter headings.
- Visual boundaries, Currier A/B textual boundaries, scribal-hand boundaries, bifolium relationships and quire structure do not always align perfectly.
- Some pages sit awkwardly between “astronomical” and “cosmological”.
- Human figures occur in more than one visual regime.
- Labels, running text, diagrams and star-marked entries can behave differently even inside one conventional section.
- The safest model is not “six independent books bound together”, but a codex containing several overlapping systems of organisation.
- A section label is useful when it tells us where to look. It becomes dangerous when it tells us what the unread text must mean.
Why We Need Sections at All
Imagine opening more than two hundred surfaces of unread writing without any descriptive categories.
You need handles.
Researchers therefore group folios by obvious visual differences. A page dominated by one plant-like drawing belongs with other plant-dominated pages. A circular diagram with a ram or bull belongs with zodiac-like material. Small human figures standing in coloured pools belong with other pages using similar imagery. Rows of plant fragments and vessels form another recognisable family. Short paragraphs marked by stars form another.
This is good scholarship.
Classification reduces complexity.
The problem begins one step later.
Once a plant page is called “herbal”, its text starts to feel botanical.
Once a page with vessels is called “pharmaceutical”, the labels start to feel like ingredient names.
Once star-marked entries are called “recipes”, their syntax starts to feel procedural.
The section name begins doing interpretive work that the manuscript has not yet earned.
The Six Familiar Labels
1. Herbal
The largest visual family contains pages dominated by plant-like drawings and surrounding text. The label is sensible because the page architecture resembles European herbal traditions.
What it safely tells us:
- plant-like imagery is central;
- the image and text were designed to coexist on the page;
- some plants or plant parts may recur internally;
- the visual regime differs from the later circular, human-figure and vessel-heavy pages.
What it does not safely tell us:
- that every text block is a species description;
- that every plant is medicinal;
- that every image represents one biological species;
- that the section formed one original chapter.
For the dedicated treatment, see The Herbal Pages.
2. Astronomical and Astrological
This family contains circular diagrams, stars, radial writing, familiar zodiac emblems and other celestial-looking forms.
Even here the boundary is not perfectly clean. Researchers often separate the zodiac sequence as “astrological” while reserving “astronomical” for other circular diagrams. Some folios around 67 and 68 are classified differently from neighbouring circular pages because their geometry appears to change function.
That ambiguity is useful. It tells us the visual world is more continuous than a six-folder filing cabinet suggests.
3. Balneological or Biological
These pages contain numerous small human figures, many apparently feminine, interacting with coloured pools, vessels, tubes, channels and compartments.
“Balneological” points toward historical bathing traditions. “Biological” points toward bodies and organic-looking structures.
The fact that researchers use two different modern names for the same broad region should make us pause.
The pages are visually coherent enough to group.
Their subject is not coherent enough to name with certainty.
See The Human Figures, Pools and Tubes.
4. Cosmological
“Cosmological” is often used for circular diagrams whose geometry feels larger, more abstract or less obviously astronomical. The famous Rosettes foldout sits here, along with several other unusual circular pages.
But researchers themselves note that the line between astronomical and cosmological can be hard to draw.
That is not a failure of classification.
It is evidence that our categories are approximations.
A diagram can be astronomical and cosmological at the same time. It can organise celestial bodies while expressing a model of the world. Medieval knowledge did not promise to respect our modern departmental boxes.
5. Pharmaceutical
The pages conventionally called pharmaceutical contain smaller plant fragments, repeated vessel-like forms, short labels and paragraphs.
“Pharmaceutical” is plausible because medieval materia medica could involve plant parts, containers, preparation, storage and remedies.
Yet the strongest immediate observation is representational rather than medical:
The manuscript shifts from whole plant-like organisms toward smaller addressable components.
That could support ingredients, storage, retrieval, classification or preparation. The page architecture is real. The exact medieval job remains unresolved.
See The Vessels and Plant Fragments.
6. Recipes or Starred Text
The final large textual regime contains many short entries accompanied by star-like marginal marks.
Historical recipe collections can use short entries and marginal markers, which is why “recipes” is a reasonable analogy.
But one star does not always equal one paragraph, and the text has not been translated into instructions.
The safer name is sometimes the less exciting one:
star-marked short-text entries.
See The Starred Pages.
Now Add Currier A and B
The six visual sections would be easier to trust as semantic divisions if the text divided the same way.
It does not.
Prescott Currier observed two strong textual varieties now called Currier A and Currier B. Their differences involve glyph and token distributions rather than decoded meaning.
These textual regimes correlate imperfectly with visual sections.
That matters.
If the six visual categories were six clean subject chapters, we might expect textual populations to switch neatly with them.
Instead, the manuscript behaves more like overlapping layers:
- visual regime;
- textual regime;
- scribal hand;
- quire and bifolium;
- label versus running text;
- local vocabulary;
- line and paragraph position.
None of those layers is guaranteed to be the “true chapter system”.
For the deeper textual distinction, see Currier A and B.
Now Add Scribal Hands
Lisa Fagin Davis’s influential palaeographic work proposed five main scribal hands in the manuscript. Later criticism has argued that variation may be more continuous than five sharply separated scribes suggest.
Regardless of which model eventually proves best, the important architectural point survives:
handwriting variation introduces another organisational layer that does not simply collapse into the six visual labels.
A hand can participate in more than one visual family.
A visual family can involve more than one hand.
That is exactly what we would expect from a collaborative or staged production process.
See The Scribe Problem.
Now Add the Bifolium
The physical codex introduces another complication.
A modern section is usually imagined as a sequence of pages.
A medieval codex is built from folded sheets.
Two leaves that appear separated in current reading order can belong to the same original bifolium. Conversely, facing pages can become neighbours because of assembly or rebinding rather than because a medieval author intended them as a conceptual pair.
The new specialist article The Bifolium as a Production Unit owns this physical perspective.
Once physical sheet relationships are included, “section boundary” becomes even less like a clean modern chapter divider.
Now Add Production Order
There is one more layer.
Drawing order.
Writing order.
Paint order.
Assembly order.
These need not be identical.
The new Order of Making article owns that problem.
A visual section may therefore reflect not only subject, but production workflow.
The Better Model: Overlapping Organisations
Instead of imagining one master partition, imagine several transparent maps laid over the same manuscript.
Visual map
Plants, circles, bodies, vessels, stars.
Textual map
Currier A/B, local vocabulary, line effects, labels, running text, entry-like text.
Scribal map
Differences in ductus, glyph construction and hand.
Physical map
Bifolia, gatherings, foldouts, missing leaves, later binding.
Production map
Drawing, writing, colour, assembly and later additions.
The deepest question is not which map is “correct”.
All may be correct at their own level.
The mistake is expecting one organisational layer to explain all the others.
Could the Six Sections Still Represent Six Subjects?
Yes.
Nothing in this article proves that the visual groups are semantically meaningless.
They may correspond strongly to subject matter.
The point is narrower:
we have not deciphered enough text to know whether the boundaries are:
- subjects;
- document genres;
- work stages;
- source traditions;
- workshop divisions;
- different intended uses;
- or some combination.
The six-section convention remains an excellent map.
It should not be mistaken for the territory.
Primary School: Sort by What You Can See
Give a child twelve unfamiliar pictures from the manuscript and ask them to sort the pictures into groups.
Then ask a second question:
Does sorting tell you what the words say?
The child quickly discovers the difference between classification and translation.
Secondary School: Compare Three Maps of One Manuscript
Make three columns:
- visual section;
- Currier type;
- scribal hand.
Then place sample folios into the table.
The student sees immediately that one manuscript can have several valid classifications at once.
This is a powerful lesson far beyond Voynich: classifications answer questions. Different questions create different partitions.
JC and Adult Readers: Ask What Variable Created the Cluster
A mature analysis asks why a group exists.
Did pages cluster because they share subject?
Because the same scribe wrote them?
Because one exemplar supplied their images?
Because one quire was produced in one work session?
Because a particular page format requires a particular vocabulary?
These explanations can produce similar clusters while implying very different manuscript histories.
Reader Checklist: Before You Treat a Voynich Section as a Chapter
- Is the section name modern or written in the manuscript?
- Is the category based on image, text, hand, quire or binding?
- Does the boundary align with Currier A/B?
- Does the boundary align with scribal hand?
- Does the boundary align with bifolium structure?
- Does a page near the boundary look transitional?
- Are there figures or motifs shared with another section?
- Does the same textual vocabulary cross the boundary?
- Could the grouping reflect workflow rather than subject?
- What evidence would distinguish those explanations?
Frequently Asked Questions
Are the six Voynich sections wrong?
No. They are extremely useful descriptive categories. The caution is that they are not six deciphered medieval chapter headings.
Does Yale use six sections?
Yes. Yale’s public Beinecke description presents six distinct illustration groups for reader orientation. That is a sensible catalogue-level summary of an undeciphered manuscript.
Why can astronomical and cosmological pages be hard to separate?
Because circular celestial-looking diagrams can represent several overlapping intellectual jobs. Geometry alone may not reveal whether the intended emphasis is astronomical observation, astrological organisation, cosmology, calendar, medicine or another classificatory system.
Could visual sections correspond to original source manuscripts?
Possibly. A compiler could combine material from different exemplars. But that is a production hypothesis requiring evidence; visual difference by itself does not prove multiple sources.
Could they correspond to different authors?
Possibly, but scribal-hand variation, authorship and source authorship are separate concepts. One scribe can copy several authors; several scribes can copy one work.
Research Foundations
- Yale Beinecke Library — The Beinecke Cipher (Voynich) Manuscript
- René Zandbergen — Voynich illustrations and section classifications
- Voynich — eduKateSG Master Article
The Final Idea
The six sections are useful because they make the manuscript easier to enter.
They become dangerous only when we forget who named them.
We did.
Not the people who made the manuscript.
The better model is a book with several overlapping organisations—visual, textual, physical, scribal and procedural—whose relationships remain partly recoverable even while the words remain unread.
The map is useful. The mistake is confusing it with the manuscript’s own map of itself.
Continue Through the Voynich Research Map
Return to the canonical master for the entire evidence architecture, or continue into the specialist visual, textual and production nodes.