Voynich and Padua | The Star-Marker Problem: A Bullet Point Is Not a Recipe

Quire 20 looks like a list of recipes because hundreds of short paragraphs are marked by stars.

That visual rhythm is powerful.

One star.

One entry.

Next star.

Next entry.

The temptation is to treat the marker as genre evidence.

But a paragraph marker can tell us where an entry begins without telling us what the entry is.

Quick Read

• Voynich Quire 20 contains hundreds of marginal star-like markers associated with short text entries.
• René Zandbergen’s codicological overview counts 324 surviving marginal stars in the section, with at least one bifolium missing.
• The section is conventionally called “recipes” because the short-entry format resembles practical recipe collections.
• Known fifteenth-century recipe manuscripts often separate entries using headings, rubrication, paragraph initials, contents lists and other visual devices.
• Wellcome MS.5262, from the first quarter of the fifteenth century, contains a contents list for 129 recipes.
• British Library recipe collections likewise use rubrics, display script and entry-level formatting.
• Therefore segmentation is expected in recipe books—but the specific star marker does not independently prove the entries are recipes.
• The star system may function as bullets, category markers, cross-references, celestial symbols or a thematic navigational layer.
• The high-value test is whether star type, colour, tail, position or associated text correlates with entry function.

What the Stars Actually Tell Us

The safest observation is structural.

Quire 20 contains short entries.

Marginal stars are repeatedly associated with them.

That creates segmentation.

Segmentation is real even if genre remains unknown.

star → entry boundary is much safer than star → recipe.

Why “Recipe Section” Is Understandable

The conventional label is not irrational.

The manuscript has already shown plants, vessels and body-like imagery.

At the end comes a long list of short entries.

Known medical books often end up containing compact practical recipes.

The overall architecture makes the interpretation plausible.

But architecture is not translation.

Real Recipe Books Make Entry Boundaries Visible

Wellcome MS.5262 is a fifteenth-century recipe collection from Worcestershire, produced in the first quarter of the century.

Its contents list records 129 recipes.

The point is simple.

A user needs to know where one remedy stops and another begins.

Visual segmentation is operational infrastructure.

Recipe Manuscripts Use Many Segmentation Technologies

Surviving recipe manuscripts separate entries in different ways.

• rubrics;
• larger display script;
• coloured initials;
• paragraph marks;
• headings naming disease or remedy;
• blank space;
• contents lists;
• marginal marks.

This diversity matters.

A star-shaped bullet is not required for recipe culture.

And a star-shaped bullet can theoretically mark non-recipe material too.

The Marker Can Be Semantically Empty

A modern bullet point tells us:

new item.

It does not tell us whether the item is:

• a recipe;
• a law;
• a shopping item;
• a definition;
• a research result.

The same principle may apply to Quire 20.

The star could be purely structural.

But the Stars Are Not All Identical

Voynich’s marginal stars vary.

They can have different point counts.

Some are coloured.

Some have tails.

Some do not.

This creates a high-value question:

is star variation decorative noise, scribal variation or encoded metadata?

If star classes correlate with text classes, the marker system is semantically richer than a bullet point.

The Star–Paragraph Alignment Is Not Perfect

Zandbergen’s Quire 20 description notes pages where star count and obvious paragraph count do not align perfectly.

Some pages include stretches of running text while stars continue in the margin.

This matters because a simplistic one-star-one-recipe model should predict clean alignment.

Misalignment may reflect:

• lost or uncertain paragraph boundaries;
• stars added at a separate production stage;
• markers referring to lines rather than paragraphs;
• subentries;
• semantic classes independent of text segmentation.

The Zodiac Connection Makes the Marker More Interesting

Similar star forms appear elsewhere in Voynich, especially in celestial and zodiac-associated contexts.

This creates another possibility.

The star may not be merely generic punctuation.

It may deliberately connect Quire 20 to a manuscript-wide symbolic system.

That hypothesis must be tested structurally:

• Do star subtypes recur across sections?
• Do associated labels share vocabulary?
• Do colours correspond?
• Do point counts matter?
• Do tails correlate with text position?

What a Real Recipe Should Contain

Segmentation alone is not enough.

An executable recipe normally needs some combination of:

• indication or disease;
• ingredient;
• quantity;
• preparation action;
• administration;
• timing;
• dosage;
• storage.

Not every medieval recipe states every field.

But across hundreds of entries, recurrent operational slots should appear.

That is the real genre test.

Formula Position Can Be Tested Without Translation

If Quire 20 entries are recipes, certain word families may prefer:

• entry beginnings;
• entry endings;
• second position;
• positions near repeated quantity-like tokens.

This can be compared with known recipe syntax at the level of structural slots rather than word meaning.

A strong model should predict held-out entries.

The Contents-List Control Is Brutal

MS.5262 explicitly records a contents list for 129 recipes.

Other medical commonplace books also expose tables or indexes.

Voynich Quire 20 has no obvious familiar index.

If it contains hundreds of practical entries, how does a user retrieve one?

Possible answers include:

• star subtype;
• entry order;
• memorised sequence;
• missing index leaves;
• cross-reference words;
• visual scanning by opening formula.

Each can be tested.

Padua Does Not Get the Recipe Section for Free

Paduan pharmacy, Montagnana’s Antidotarium and working receptaria make recipe culture historically plausible.

But fifteenth-century recipe collections existed across Europe.

Therefore short practical entries do not localise Padua.

Only Paduan-specific textual descent, measurement systems, local ingredient vocabulary or provenance could do that work.

The Star-Marker Test

1. Classify every star by points, colour, centre and tail.
2. Map each star to paragraph boundary and line position.
3. Test whether star class predicts entry vocabulary.
4. Compare stars in Quire 20 with stars in zodiac and other sections.
5. Measure recurrent word positions inside entries.
6. Search for quantity-like and action-like structural slots.
7. Compare against actual fifteenth-century recipe collections.
8. Keep “recipe” as provisional until operational structure emerges.

What the Stars Can Legitimately Support

• Quire 20 is strongly segmented.
• The markers likely participate in navigation or entry organisation.
• Recipe collections are a plausible comparator because they also require strong segmentation.
• Star variation may carry metadata and deserves formal testing.

What They Cannot Support

• that every starred entry is a recipe;
• that the stars are celestial symbols;
• that the star shape is Paduan;
• that one star equals one complete entry in every case;
• that the section’s traditional name is its actual genre.

Research Sources

• René Zandbergen — Voynich Quire 20
• Wellcome Collection — Fifteenth-Century Medical Recipe Collection, MS.5262
• British Library — Royal MS 12 G IV, Medical Tracts and Recipes
• Voynich — The Starred Pages

The Final Idea

The stars almost certainly organise something.

That is already useful.

But a bullet point is not a genre. Quire 20 becomes a recipe section only when its internal structure begins to behave like recipes, not merely because the margin looks like a list.