Voynich Q-Series: Why qo- Begins So Many Words

One Voynich character behaves as if it knows where the left edge of a token is.

In EVA, researchers call it q.

That modern letter is only a transcription label.

It does not mean the Voynich sign sounds like Latin q.

It does not prove that the following sign is a vowel.

And yet the resemblance to familiar q-u spelling is almost impossible for a modern reader to ignore because Voynich q behaves in a remarkably constrained way.

It appears overwhelmingly at the beginnings of space-delimited tokens.

When it begins a token, it is essentially always followed by the sign conventionally transliterated o.

So researchers routinely talk about the qo- family as though q and o form a natural unit.

Across the manuscript, roughly fourteen per cent of tokens begin with q under one widely used classification.

But the proportion is not stable.

In some Currier-B-heavy material it becomes far more common.

In other textual regimes it falls.

And in label populations—one of the strongest specialised registers in the manuscript—q-initial forms become conspicuously rare.

This is exactly the kind of behaviour that tempts us to assign grammar.

Maybe qo- is a prefix.

Maybe it marks a verb.

Maybe it means “the”.

Maybe it turns a label-like noun into a prose form.

Maybe q and o are not two independent signs at all.

Maybe q triggers an encoding state whose first visible consequence is o.

All of those are mechanism hypotheses.

The observation is narrower and stronger.

The Voynich writing system has a highly restricted q→o opening construction whose use depends on textual context.

That construction deserves its own article because it sits at the intersection of nearly every text problem we have already built.

Word boundaries.

Word families.

Entropy.

Labels.

Currier A and B.

And eventually, syntax.

Quick Read

One-sentence answer: the Voynich sign transliterated q is an exceptionally position-sensitive symbol that occurs almost entirely at token beginnings and is essentially always followed by o, producing a qo- family whose frequency changes sharply by Currier regime and document role; this makes qo- one of the clearest structural prefixes or prefix-like constructions in the visible script, while leaving open whether it is grammatical, orthographic, graphemic, cryptographic or generative.

EVA q is a transcription code, not a known sound value.
Voynich q is overwhelmingly token-initial.
Word-initial q is essentially always followed by EVA o, so q-initial and qo-initial categories are practically equivalent in many corpus descriptions.
Across one modern classification, about 14% of all tokens begin q.
The proportion varies substantially by textual regime.
It can approach one quarter of tokens in strongly Currier-B-like Quire 13 material.
It is much less frequent in some cosmological/C-like and herbal populations.
Labels conspicuously avoid q-initial forms even though q is common in running prose.
This label/prose contrast makes document role an important variable.
Common qo- forms include qok- and qot-like families that have close visible relatives without initial q.
That relationship makes prefix or detachable-element models plausible.
But q+o may instead be a compound grapheme or one orthographic unit represented analytically by EVA.
A cipher or notation can also place state-setting material at token starts.
The sequence does not prove Latin, Romance, or any language merely because Latin q is commonly followed by u.
A successful theory should predict when q appears, when it is forbidden, and what changes in a token or context when q is added.

The q-series is therefore a gift to decipherment.

It is unusually constrained.

Constrained phenomena are testable.

First: q Is Not the Latin Letter Q

EVA uses familiar Roman letters because they are easy to type.

That convenience creates one of the most persistent cognitive traps in Voynich research.

We see q.

Then o.

Our alphabetic experience whispers:

q followed by a vowel.

But EVA q and EVA o are shape labels.

The manuscript itself never tells us that one is consonantal and the other vocalic.

The visual sequence may represent:

two graphemes;
one compound grapheme;
a prefix plus a core element;
an abbreviation;
an encoding instruction plus payload;
a decorative or orthographic construction.

The similarity to familiar q behaviour is interesting only after those alternatives are controlled.

Transcription should help us refer to the sign.

It should not smuggle in its phonology.

q Is Almost a Boundary Detector

The q sign is striking because its position is so restricted.

It strongly prefers the beginning of a space-delimited unit.

This creates a natural thought experiment.

Erase the spaces from one line.

Can q help recover some token boundaries?

Often, q would nominate a likely start.

That does not prove the token is one lexical word.

It does show that q participates in the same boundary architecture as the spaces.

This is important because independent clues to segmentation are rare in undeciphered scripts.

If visible gaps and q placement agree strongly, the original writer likely recognised the same structural boundary at both levels.

Again, we know where a unit starts before we know what kind of unit it is.

q Almost Always Brings o With It

Position is only the first constraint.

The next symbol is constrained too.

In ordinary corpus descriptions, q-initial Voynich tokens are essentially qo-initial tokens because q is almost invariably followed by EVA o.

This creates one of the strongest local transition probabilities in the visible script.

Why would a writing system do this?

Several mechanism families immediately appear.

Two-letter sequence

q and o are independent graphemes, but orthographic rules make the pair nearly obligatory.

Compound grapheme

The visible pair is functionally one written unit that EVA decomposes analytically.

Prefix-plus-core

q modifies an o-initial family and therefore normally appears only before o.

Encoding state

q initiates a state whose first visible emitted component is o.

Generation template

The token grammar simply allows q only in one opening slot and requires o next.

All five can generate the surface observation.

We need additional distributions to distinguish them.

The q Frequency Changes Across the Manuscript

If q were merely a universal decorative flourish at every word start, we might expect roughly similar use across textual domains.

We do not see that.

One modern extension of Currier classification reports that roughly fourteen per cent of all tokens begin q overall.

Yet strongly B-like Quire 13 material can approach twenty-five per cent.

Some cosmological/C-like pages have much lower q rates.

Herbal material is lower than the Quire 13 extreme.

This makes q part of the manuscript’s state variation.

A theory must therefore explain two things at once:

Why q is so constrained inside a token.
Why the probability of using the q construction changes between textual regimes.

A simple universal prefix theory may be too weak unless the grammatical category itself changes by genre or dialect.

An encoding-state theory may explain frequency change naturally if states differ between Currier regimes.

A scribal theory may explain it if writers have different habits.

The distribution forces the mechanism to become specific.

Labels Avoid q

The label population gives us one of the strongest cross-register clues.

Labels do not simply sample the same word population as running prose.

Among their differences, q-initial forms are conspicuously uncommon.

This matters because several explanations make different predictions.

If q is grammatical

Labels may omit the grammatical environment in which q normally occurs.

If q marks prose mode

The writing system may switch mode for isolated labels.

If q is phonetic

The lexical classes used as labels may simply begin less often with the represented sound or syllable.

If q is generated mechanically

The generator must have a label state that suppresses q.

In every case, the label contrast turns q from a mere character curiosity into a document-role marker candidate.

The manuscript appears to know when it is writing a label.

q helps us see that.

qok- and qot- Are Especially Useful Families

Many frequent qo- tokens continue into familiar qok- and qot-like constructions.

This matters because related forms exist without the initial q.

For example, one can compare qok- forms with ok- forms and qot- forms with ot- forms.

The visible relation invites a detachable-prefix model.

q + ok-family → qok-family

If q is a genuine prefix, the q-bearing and non-q forms should have a principled relationship.

They may share a lexical stem and differ grammatically.

They may represent two encoded states of one unit.

Or their visual similarity may be an artefact of a slot grammar in which q simply occupies the outermost slot.

The crucial next question is contextual.

Do qok- and ok- forms occur in systematically different environments?

That is how a visible prefix becomes a functional prefix.

The Strange Case of qol

One useful warning comes from a form that looks as if it should be common.

q is common at beginnings.

ol-like endings and tokens are common.

Yet qol is relatively rare overall and is heavily concentrated in particular B-like material.

This tells us that Voynich word construction is not a free combination of common pieces.

There are selection rules.

A simple slot model in which every legal prefix combines freely with every legal suffix will overgenerate.

The absence or rarity of an apparently easy combination can be as informative as the abundance of another.

A writing system is defined partly by the forms it refuses to make.

qol therefore acts as a negative constraint on simplistic compositional theories.

Could q Be a Grammatical Prefix?

This is one of the most natural interpretations.

A prefix appears at the beginning of words.

It can attach to recurring stems.

It can be absent in labels if labels disproportionately contain bare names or noun forms.

It can vary by dialect or genre.

All of that fits qo- superficially.

But a grammatical prefix should do grammatical work.

If q marks a tense, case, article, preposition-like element or derivation, adding q should change the syntactic distribution of the stem in predictable ways.

For example:

q-bearing forms should prefer particular preceding classes;
they may co-occur with particular endings;
their label avoidance should match the proposed grammar;
the same q operation should generalise across many stems.

Without those relationships, “prefix” remains a positional description rather than a decoded grammatical function.

Could qo Be One Grapheme?

The near-obligatory q→o sequence raises a more basic possibility.

Perhaps our segmentation is wrong.

If qo behaves so tightly that q scarcely exists independently, perhaps the medieval writing system treats the pair as one grapheme or ligature-like unit.

Under this model, EVA has analytically split one functional sign into two characters.

This would change:

alphabet size;
token length;
bigram entropy;
word-family edit distance;
morphological parses.

The model has to answer a palaeographic question.

Does the physical writing support treating q and o as one stable stroke complex?

If q is visibly separable and behaves independently in rare contexts, a strict one-grapheme model weakens.

Again, statistics can nominate the hypothesis.

Ink must help decide it.

Could q Mark an Encoding State?

An encoded system can use token-initial material to change how the rest of the group is read.

A shift sign in another technology tells the system:

interpret what follows under state B.

q could in principle perform an analogous role.

This would explain why it lives at boundaries and why its frequency can change between textual regimes.

It might also explain q-bearing and non-q word-family pairs.

But the encoding theory needs an actual transformation.

What state does q select?

How does o participate?

What happens to the underlying plaintext?

Why are labels q-poor?

A generic “shift marker” analogy is not enough.

The rule should generate the observed distributions from readable input.

Could q Be a Scribal or Orthographic Convention?

A visible prefix-like sign need not represent separate meaning.

It may change how a following sign is written.

Historical scripts contain contextual forms.

A character can look different at the beginning of a word.

A scribe can add an entry stroke or linking flourish.

Under such a model, qo might be related to a bare o-family form without q carrying independent semantics.

The strongest test is substitutability.

If qok- and ok- forms occupy the same contexts and a future decipherment maps them to the same underlying units, orthographic variation becomes compelling.

If their contexts differ systematically, q probably carries more functional information.

Could q Be a Generation-Slot Rule?

Constrained generation explains qo- easily.

Define an optional opening slot.

If the slot chooses q, force o next.

Then continue into a limited core family.

The result naturally produces:

token-initial q;
near-obligatory qo;
qok/qot families;
low conditional entropy.

The challenge is not producing the q surface pattern.

The challenge is explaining why q use changes with Currier regime and disappears from labels in a purposeful way.

A generator that contains a prose state and a label state can do that.

But now the generator has become document-aware.

The distinction between meaningless filler and specialised notation becomes less simple.

q Is a Major Contributor to Predictability

The entropy article asked why the next Voynich character is often so easy to predict.

q gives us one concrete answer.

If q appears, o is overwhelmingly likely next.

That collapses uncertainty locally.

If q is itself restricted to token beginnings, knowing token position reduces uncertainty again.

If Currier regime changes q frequency, knowing regime changes the prior probability once more.

This shows how Voynich low entropy can arise from stacked constraints:

boundary position;
q→o pairing;
family selection;
regime state.

Entropy is not one mysterious global property.

It is the sum of many local restrictions like this one.

q Can Help Test the Word-Boundary Hypothesis

If q truly marks token beginnings, its placement can be compared with ambiguous spaces.

Suppose a disputed gap appears before q.

The ordinary q distribution makes a genuine boundary more plausible.

Suppose a gap is inserted inside qo.

The near-obligatory pairing makes that segmentation suspicious.

This does not let q decide every space.

It lets character distribution become an independent clue to tokenisation.

The right segmentation should make q behaviour more coherent, not less.

q Can Help Test Label Grammar

Why do labels avoid q?

This is one of the cleanest questions a future decipherment can answer.

If q becomes a grammatical article-like or prepositional element, the absence may make immediate sense.

Labels may use bare lexical forms.

If q is an encoding state used for prose, label mode may explain the suppression.

If q is phonetic, the decipherment should show that label vocabulary simply contains fewer q-class underlying forms.

Each solution predicts a different reason for the same distribution.

The q-label contrast is therefore a valuable hostile test for any proposed function.

q Can Help Test Currier Transformations

Suppose Currier A and B are two orthographic states of one underlying system.

The q rate changes substantially between them.

A transformation model should explain that difference.

Perhaps one regime expresses a grammatical element overtly that another omits.

Perhaps one encoding state uses q for a class that another encodes differently.

Perhaps q frequency is downstream of different vocabulary rather than different grammar.

A good A↔B model should make q frequencies an expected consequence.

If it leaves q as an unexplained exception, it has not yet captured one of the largest visible regime differences.

The q-Series Is Not Proof of Romance

Because q is visually reminiscent of familiar alphabetic q and is nearly always followed by one sign, it is tempting to search languages with strong q-vowel orthography.

This reasoning is contaminated by EVA.

If René Zandbergen and Gabriel Landini had assigned the sign the letter x instead, we would speak about xo- and the Romance intuition would weaken immediately.

The underlying manuscript pattern would be identical.

This is an excellent demonstration of why transliteration labels should never be allowed to seed language identification unconsciously.

Language must emerge from systematic mappings across the corpus, not from the mnemonic convenience of EVA letters.

The q-Series Is Not Proof of a Prefix Either

Calling q a prefix is less dangerous than assigning it a meaning, but still partly theoretical.

Positionally, it behaves prefix-like.

Functionally, we do not yet know whether it is detachable in the linguistic sense.

A graphical compound can occupy the same visible position.

A state marker can too.

The safest wording is:

q is a highly restricted token-initial component that participates in a qo- opening construction.

That sentence contains everything the manuscript currently earns.

What Survives the Q-Series Work

q is one of the most position-sensitive common signs in Voynichese.
It overwhelmingly marks token beginnings.
q is essentially always followed by o in ordinary word-initial use.
qo- participates in dense visible word families.
q frequency varies strongly across textual regimes.
Strongly B-like Quire 13 material is especially q-rich.
Labels strongly suppress q-initial forms.
q therefore interacts with both textual regime and document role.
The q→o rule contributes materially to low conditional entropy.
A successful writing-system model should explain q, o, q-bearing families and label avoidance as one system.

What Does Not Survive as Established Knowledge

EVA q is a phonetic /q/.
EVA o is a vowel.
qo corresponds to Latin qu.
q proves a Romance language.
q is a decoded grammatical prefix.
q is a proven article, preposition, tense marker or case marker.
qo is proven to be two independent graphemes.
qo is proven to be one grapheme.
q is a proven cipher-state marker.
q-rich pages have one decoded semantic topic.

The distribution is strong.

The semantic label remains open.

A Better Q-Series Analysis

Keep EVA labels phonologically neutral.
Measure q only after explicit tokenisation.
Separate q, qo and whole q-bearing families.
Compare q-bearing forms with their closest non-q relatives.
Condition on Currier regime.
Condition on label versus prose role.
Condition on line and paragraph position.
Test one-grapheme and multi-grapheme qo representations.
Require any grammatical function to predict surrounding token classes.
Require any encoding function to generate q distributions from plausible underlying text.

What Would Count as a Real q Breakthrough?

Imagine a decipherment built without using q as an assumed prefix.

It independently recovers an underlying grammatical system.

Across hundreds of forms, adding q to an o-initial family corresponds to one stable grammatical operation.

The operation predicts why labels avoid q.

It predicts the q-rate difference between A and B.

And it works on unseen pages.

That would move q from positional prefix to grammatical prefix.

Or imagine palaeographic analysis shows qo is consistently executed as one inseparable sign, and treating it as one grapheme removes several otherwise unexplained entropy anomalies while improving decipherment.

That would move it toward a compound-grapheme interpretation.

Or a historically plausible encoding procedure could show why q initiates one state, why o follows, and why the state is suppressed in labels.

That would make it cryptographic.

The point is the same:

a q theory becomes strong when it predicts all the places q does not occur as well as the places it does.

Primary School: The Rule at the Beginning

Create an invented writing game.

Tell the child that symbol △ may appear only at the beginning of a group, and whenever △ appears, ○ must come next.

Ask the child to spot illegal strings.

△○AB — allowed.
A△○B — not allowed.
△AB — not allowed.

The child understands the Voynich q observation immediately without assigning any meaning to △.

Structure can be learned before semantics.

Lower Secondary: Same Core, With and Without q

Give students pairs such as:

OKA / QOKA
OTE / QOTE
OKY / QOKY

Ask for four possible explanations.

q is a grammatical prefix;
q is a spelling variant;
q is an encoding marker;
q is a generation slot.

Then ask what extra evidence would separate them.

The learner moves from resemblance to mechanism.

Upper Secondary: Use Labels as the Crossing Case

Suppose q appears in twenty per cent of prose tokens but almost never in labels.

Ask students what each theory predicts.

If q is an article-like grammatical element, label avoidance may be expected.

If q is a universal phoneme, the avoidance requires a lexical explanation.

If q is a prose-mode code marker, the result is direct.

The student learns why crossing text roles is more informative than counting q globally.

JC and Adult Readers: q Is a Conditional Probability Problem

At a higher level, q is interesting because several conditional probabilities are extreme.

P(o | q) is very high.

P(q | token-initial) is much higher than P(q | internal).

P(q | Currier B, Quire 13) is higher than in several other regimes.

P(q | label) is low.

A mechanistic model should explain the whole conditional table rather than one marginal frequency.

That is the mature q question:

what hidden variable makes all of these conditional probabilities emerge together?

A Parent and Teacher Guide

Keep EVA q separate from Latin q.
Notice position before guessing meaning.
Notice that q predicts o.
Compare q-rich and q-poor text types.
Use labels as an independent register.
Compare q-bearing forms with non-q relatives.
Require one mechanism to explain both presence and absence.

The broader reasoning lesson is simple:

the most informative rule is often not what an element resembles, but the conditions under which it is allowed to appear.

Reader Checklist: Before You Explain qo-

Are you unconsciously pronouncing EVA q and o?
Are q and o assumed to be independent graphemes?
Is q genuinely token-initial in the corpus used?
Are uncertain spaces affecting the count?
How often does q occur without o?
How does q frequency change by Currier regime?
How does it change between labels and prose?
Do q-bearing forms have non-q family counterparts?
Does adding q change contextual distribution?
Does the theory explain rare combinations such as qol?
Can the mechanism reproduce the low entropy created by q→o?
What would make the proposed q function fail?

Frequently Asked Questions

What is the Voynich q-series?

It is the large family of tokens beginning with the Voynich sign transliterated q, which is essentially always followed by EVA o and then continues into common families such as qok- and qot-.

Does Voynich q sound like q?

Unknown. EVA letters are transcription labels chosen for convenience, not decoded phonetic values.

Is q always followed by o?

In normal word-initial use, essentially yes, with rare or exceptional transcription cases not changing the broad rule.

How common are q-initial tokens?

One modern corpus classification reports about fourteen per cent overall, with large variation across textual regimes and much higher rates in some strongly Currier-B-like material.

Do labels use q?

They use q-initial forms much less frequently than ordinary running text, one of several ways in which labels form a specialised register.

Is q a prefix?

It is positionally prefix-like. Its grammatical, phonological or encoded function has not been established.

Could qo be one character?

It is a plausible representation question because the sequence is so tightly linked. Palaeographic and predictive evidence would be needed to establish one-grapheme status.

Why does q matter for entropy?

Because once q appears, the next character is highly predictable, and q itself is strongly predictable from token position and textual regime.

What is the strongest current conclusion?

q is a highly constrained opening component whose qo- construction changes systematically across textual regimes and document roles. Any serious Voynich writing-system model must explain that entire distribution.

Research and Further Reading

The Final Idea

q is one of the manuscript’s best small mysteries because it is almost too well behaved.

It waits at the beginning.

It brings o with it.

It grows into familiar token families.

It becomes common in some textual states.

It nearly disappears when the manuscript switches into label mode.

Those facts are stronger than any attractive translation we can currently attach to q.

They tell us that the sign is embedded in the operating rules of Voynich writing.

Whatever q means—if “meaning” is even the right level—it is not free.

It has permissions.

It has prohibitions.

It has contexts.

And in an undeciphered manuscript, a rule about where something may exist is often the beginning of learning what kind of thing it is.

Voynich | Everything eduKate Knows and Tested | The Q-Series: Why qo- Lives at the Beginning of So Many Voynich Words