Every undeciphered manuscript invites somebody to look for the key.
The Voynich Manuscript makes the temptation almost unbearable because it contains sequences that look as though they ought to explain something.
A vertical list of single characters.
Numbers 1 to 5 beside another sequence.
A circular ring in which seventeen character-units are repeated four times.
Marginal strings that do not look like ordinary running prose.
Researchers have long used a convenient descriptive phrase for such material:
key-like sequences.
The hyphen matters.
They look key-like.
They are not proven keys.
This article is about what these abnormal sequences can tell us before we turn them into the answer we want them to be.
Quick Read
- The Voynich Manuscript contains several sequences of isolated characters or short words outside ordinary paragraph prose.
- These are often described as “key-like sequences” in technical cataloguing because their form resembles inventories, alphabets, ordered lists or cipher keys.
- The term does not establish that any sequence is cryptographic.
- f49v contains a vertical sequence of single Voynich characters in the left margin, with Arabic numerals 1–5 beside the upper five; both may have been added after the main text.
- f76r contains another vertical sequence of single Voynich characters aligned with some text lines.
- f66r contains marginal sequences of characters and short words, plus mixed later-looking writing near a reclining figure.
- f1r has faded or erased vertical character sequences in the right margin that have been tentatively associated with a later hand.
- f57v is different: its character and word sequences are integrated into the original circular diagram rather than obviously added as marginalia.
- One f57v ring contains a sequence of 17 character-units repeated four times with only small local differences.
- Such sequences may represent alphabets, ordered categories, diagram labels, mnemonic inventories, cipher-related structures, counting devices, later study notes or other document functions.
- A genuine “key” interpretation must explain the surrounding page and make stable predictions elsewhere.
Why “Key-Like” Is a Good Research Term
Good terminology describes without pretending to decode.
“Key-like” does exactly that.
A vertical list of isolated signs can resemble:
- an alphabet;
- a cipher key;
- a lookup table;
- a numbered index;
- a mnemonic list;
- a scribal exercise;
- a classification sequence.
Those are different mechanisms.
The shape of the list is enough to justify comparison with them.
It is not enough to choose among them.
A key-like sequence is an object class. “Cipher key” is an interpretation.
f49v: Five Numbers Beside Five Voynich Characters
f49v is one of the most seductive pages in the entire manuscript because it contains ordinary Arabic numerals beside Voynich signs.
In the left margin is a vertical sequence of single Voynich characters aligned near the beginnings of lines. Beside the upper five are the numerals 1, 2, 3, 4 and 5.
This looks wonderfully explicit.
It is also exactly where evidence discipline matters.
The numerals may have been added after the main text.
The character sequence itself may also be secondary.
Therefore the safest claim is not:
these five glyphs mean one, two, three, four and five.
It is:
a later or secondary writer appears to have ordered or indexed five Voynich characters using the numbers 1–5.
That is still valuable.
It tells us somebody interacted analytically with the script.
But indexing a character is not the same as assigning it a numerical value.
The canonical quantitative owner remains Where Are the Numbers?.
The Blackboard Test
A teacher writes:
- 1. A
- 2. B
- 3. C
- 4. D
- 5. E
No student concludes that A means one.
The numbers order the letters.
They do not necessarily define their values.
f49v requires the same caution.
f76r: Another Vertical Character Sequence
f76r contains another detached vertical sequence of single Voynich characters in the left margin, aligned with some lines of the page.
There are no helpful Arabic numerals beside it.
That makes its role even less obvious.
Possible functions include:
- line indexing;
- scribal planning;
- later study notes;
- categorical markers;
- a local sequence related to the page’s internal organisation.
The important point is not to force f76r into the same mechanism as f49v merely because both are vertical.
Visual similarity creates a comparison.
It does not yet create identity.
f66r: Characters, Short Words and a Crowded Margin
f66r is more complicated.
The left margin contains sequences of characters and short Voynich words. Near the lower left is additional writing associated with a reclining human figure and other small visual elements, including material that has been read in different ways as German-like or Latin-script writing.
This page is therefore a layered environment.
Main Voynich prose.
Voynich sequence material.
Later-looking or mixed writing.
Illustration.
A key-like interpretation must first establish which layer each mark belongs to.
Otherwise later marginal activity can be mistaken for original encoding machinery.
See Marginalia and the Later Hands.
f1r: Faded Vertical Sequences at the Beginning of the Book
f1r contains faded or erased vertical sequences of single characters in the right margin.
They have been tentatively associated with a later hand, possibly Johannes Marcus Marci.
If that attribution is correct, the marks are especially revealing because they show a later owner or reader interacting with the script itself.
But the attribution is tentative.
Again, layer before meaning.
f57v Is Different Because the Sequence Belongs to the Diagram
f57v should not be lumped casually with later marginal sequences.
Its character and word sequences are integrated into a circular diagram.
There is no obvious reason to call them extraneous.
The page contains four concentric bands of circular writing and four outward radial text items, plus labels near four central human figures.
One ring is extraordinary.
It contains a sequence of seventeen character-units repeated four times around the circle, with only minor local deviations.
This is one of the most controlled repetitive structures anywhere in the manuscript.
Whatever f57v is doing, the maker wanted the reader to encounter nearly the same ordered character sequence four times.
Why 17 Attracts Alphabet Theories
Seventeen repeated units naturally invite the idea of an alphabet or character inventory.
Historically, researchers have indeed wondered whether the ring is some kind of alphabetic sequence.
The temptation is reasonable.
The conclusion is not established.
Several problems remain:
- the units are not all obviously single elementary glyphs;
- some may be compound forms;
- the broader Voynich character inventory is larger than seventeen under many segmentations;
- rare glyphs and ligatures complicate any inventory;
- the sequence may classify states, not letters;
- the four repetitions are not perfectly identical.
The dedicated Alphabet Problem article explains why even counting Voynich characters is not trivial.
The Four Repetitions Are More Important Than the Number 17
Seventeen is attractive because it invites symbolic interpretation.
Four near-repetitions are more operationally useful.
They let us compare:
- which positions remain identical;
- which positions vary;
- whether variation follows a rule;
- whether the same compound forms recur;
- whether each cycle aligns with one of the four central figures;
- whether start markers align the sequences consistently.
This is internal evidence.
We do not need to know what the units mean to test whether the four cycles are structurally equivalent.
That is far stronger than immediately naming 17 as an alphabet size, lunar number, codebook or ritual count.
Could f57v Be an Alphabet?
Possible.
A strong alphabet interpretation would need to explain:
- why the sequence is repeated four times;
- why some units vary;
- how the seventeen units map onto the larger writing system;
- why common and rare Voynich forms are represented as they are;
- whether sequence order predicts anything elsewhere.
An alphabet interpretation should become more useful outside f57v.
If it explains only the ring itself, it remains descriptive speculation.
Could It Be a Cipher Key?
Also possible in the broadest sense.
But “cipher key” is a very strong claim.
A cryptographic key should participate in an executable mapping or transformation.
What maps to what?
Where is the corresponding plaintext or ciphertext unit?
How is the key applied?
Does it work across unseen text?
Without that mechanism, “key” is merely a metaphor for a tidy sequence.
See Cipher, Plaintext or Generated System?.
Could It Be a Mnemonic Inventory?
A repeated ordered sequence can help memory.
Four copies around a circle can organise four parallel categories.
Central figures can serve as anchors.
Labels can distinguish local states.
This makes mnemonic or pedagogical comparison reasonable.
But a mnemonic interpretation needs a repeatable relationship between position and content.
“It could help memory” is not yet a mechanism.
Could It Be a Taxonomy?
Lists do not have to be alphabets.
A sequence can order:
- categories;
- states;
- days;
- degrees;
- ingredients;
- people;
- celestial units;
- stages in a process.
The circular geometry of f57v makes ordered-category interpretations at least as important as alphabetic ones.
The central problem is still external anchoring.
What independently observable thing does each position correspond to?
Could It Be a Scribal Exercise?
A sequence of characters can be practice.
Scribes copy alphabets.
They test pens.
They rehearse difficult forms.
This possibility is more natural for marginal or later sequences such as f49v or f1r than for the carefully integrated f57v rings.
Document context matters.
A practice sequence in a margin and a sequence built into a cosmological diagram should not be assumed to share one function merely because both contain isolated signs.
The Layer Problem: Original, Contemporary or Later?
Before decoding a key-like sequence, ask when it entered the book.
There are at least three broad possibilities:
- integral production: part of the original page design;
- contemporary working mark: added during production but not part of the intended reader-facing content;
- later annotation: added by a subsequent reader, owner or researcher.
f57v strongly belongs in the first discussion because its sequences are integrated into the diagram.
f49v is much less secure as original content.
f1r may be later still.
Mixing these layers produces false universals.
The Sequence Problem Is Also a Segmentation Problem
What counts as one item?
A single visible stroke?
A compound glyph?
A ligature?
A short word?
Some f57v units that look like multiple EVA characters may function as one graphical unit.
Historical discussion of the 17-item sequence has already noted that certain combinations may need to be treated as single units if the sequence is to be interpreted structurally.
This means the apparent list length depends partly on the character model.
The EVA and Segmentation article owns that problem.
The Sequence Problem Is Also a Direction Problem
A vertical list has an obvious top and bottom.
A ring does not.
To compare the four f57v cycles position by position, we need aligned start points and a consistent direction.
Fortunately, the diagram appears to provide start markers for the concentric bands.
That makes the repeated structure far more testable than an arbitrary circular transcription.
See Which Way Does Voynich Read? for the direction, rotation and wrap-around sequence problem.
What a Real Key Interpretation Would Have to Do
- Specify which sequence is the proposed key.
- Establish that the sequence belongs to the relevant production layer.
- Define each unit consistently.
- Define sequence order and direction.
- Explain repeated and variant positions.
- Specify what the units map to.
- Show how the mapping transforms ordinary Voynich text.
- Apply it beyond the page where the key was found.
- Work on unseen text without changing values.
- Explain Currier A/B and document-role variation if the key is claimed to be manuscript-wide.
- Produce coherent output rather than isolated convenient words.
Anything less may still be a useful structural hypothesis.
It is not yet a key.
What These Sequences Do Give Us
- Evidence that Voynich characters can be used outside ordinary prose.
- Evidence of list-like and sequence-like document roles.
- A rare place where individual characters become addressable units.
- Evidence that later readers may have analysed or indexed the script.
- A highly controlled repeated character structure on f57v.
- A bridge between alphabet, numeral, diagram, marginalia and cipher questions.
- A set of internal structures that can be tested without assigning semantics.
Primary School: A List Is Not a Key
Write five unknown symbols on the board and number them 1 to 5.
Ask the class what the numbers tell us.
They tell us order.
They do not tell us what the symbols mean.
This is the central lesson of f49v.
Secondary School: Compare the Four f57v Cycles
Represent the four repeated sequences as four rows.
Align all seventeen positions.
Mark:
- identical positions;
- variant positions;
- compound forms;
- possible scribal damage;
- positions aligned with visual features.
The exercise separates structural comparison from semantic guessing.
JC and Adult Readers: Treat Each Sequence as a Document Role
Do not pool f49v, f57v, f66r and f76r simply because all contain isolated signs.
For each, record:
- physical location;
- orientation;
- relationship to main text;
- relationship to illustration;
- ink/layer evidence;
- unit type;
- repetition;
- numbering;
- possible later hand.
Then ask which sequence families really belong together.
The category “key-like” should remain a search class until the manuscript itself divides it more precisely.
Reader Checklist: Before You Call a Voynich Sequence a Key
- Is it original or later?
- Is it integrated into the page or marginal?
- What counts as one unit?
- What establishes the start point?
- What establishes direction?
- Does the sequence repeat?
- Are repetitions exact or variant?
- What does the proposed key map?
- Does the mapping work outside the sequence?
- Are numbers beside characters indexing them or assigning values?
- Does the hypothesis survive alternative glyph segmentation?
- Can another researcher reproduce the mapping?
Frequently Asked Questions
Has the Voynich cipher key been found?
No accepted cryptographic key has been established. “Key-like sequences” is a descriptive research term for unusual ordered strings of characters or short words.
Do the numbers 1–5 on f49v decode five Voynich glyphs?
Not established. They may simply number or index the adjacent characters, and both the numbers and sequence may be later than the main text.
What is special about f57v?
Its sequence material is integrated into a circular diagram. One band contains a 17-unit sequence repeated four times with small differences, making it an unusually controlled internal structure.
Could the 17 units be an alphabet?
It is a plausible hypothesis class, but not established. The unit count depends partly on glyph segmentation, and a successful alphabet interpretation must explain the rest of the writing system.
Why study these sequences if they do not translate anything?
Because they expose characters in unusual document roles—isolated, ordered, repeated and sometimes numbered—giving stronger constraints on the writing system than ordinary prose alone.
Research Foundations
- René Zandbergen — Voynich Writing System and key-like sequences
- Quire 8 — f57v repeated circular sequences
- Voynich transliteration and locus types
- The Alphabet Problem
The Final Idea
The Voynich Manuscript contains places where its script stops behaving like prose and starts behaving like an inventory.
That is rare.
It is valuable.
It does not entitle us to call every tidy sequence a cipher key.
What it gives us is better:
characters placed where their order, repetition and individuality become unusually visible.
The key-like sequences may not be the key. They are places where the manuscript exposes the machinery of its own notation more clearly than usual.