Phase 1: Comprehensive Glyph & Symbol Analysis

Introduction & Corpus Overview

• Manuscript Context: The Voynich Manuscript is a 15th-century codex (carbon-dated ~1404–1438) of mysterious origin (likely Central European). It is famously un-deciphered, containing illustrations and text organized into distinct sections (botanical, astronomical, biological/balneological, pharmaceutical, etc.).
• Script Characteristics: The text comprises approximately 170,000 characters written in an unknown script. Analysts identify about 20–30 unique glyphs (the exact count varies with how one defines distinct versus combined symbols). The text is usually transcribed using the Extensible Voynich Alphabet (EVA), which assigns Latin letters or digraphs to each Voynich glyph for analysis. The writing appears to run left-to-right in short paragraphs; attempts to read it right-to-left (as in Hebrew or Arabic) did not improve coherence, suggesting a left-to-right reading order.
• Language-Like Patterns: Statistical tests indicate the Voynich text behaves like a natural language in some respects. Word lengths follow a realistic distribution (Zipf's law) and certain repetitive structures suggest an organized grammar rather than random gibberish. At the same time, some features (like highly repetitive bigrams and unusual line-level patterns) are too regular for an ordinary known language, hinting at a possible cipher or constructed script.

Voynich Glyph Inventory & Categories

The Voynich script uses a specific set of glyphs, which we enumerate here using their EVA transliteration. Each glyph's EVA character(s) correspond to a distinct shape in the manuscript:

Common Simple Glyphs

These are the most frequent symbols, often resembling Latin lowercase letters. For example, o (a small circular shape) is the single most common character (~11.4% of all characters). Other core glyphs include e (loop shape, like a cursive "e", ~8.5%), a (loop or "alpha"-like shape, ~7.8%), i (a simple vertical stroke, ~7.5%), and n (resembles a "2"-shaped or "r"-shaped curve, ~6.7%). These five constitute roughly 40%+ of the text and appear in all sections. Notably, many of these look like Latin letters, suggesting the script's designer may have borrowed familiar shapes. Another frequently used glyph is y, transcribed for a shape that looks like a "9"-loop; y often appears at word endings (discussed later).

Gallows (Tall Glyphs)

Four less-common symbols are elongated "gallows" letters, transcribed as t, k, f, and p. These glyphs have long vertical strokes often with elaborate loops or ascenders, making them visually prominent. They occur relatively infrequently overall (each a few percent of the text) but show positional bias: gallows frequently appear at the beginnings of words and lines. In fact, certain gallows (especially t and k) occur disproportionately in line-initial positions, a phenomenon noted by Voynich researchers as a possible paragraph or section marker. For example, many paragraphs begin with a gallows character, and roughly one in seven words starts with one of these tall letters (far more than random distribution would predict). This suggests gallows glyphs may carry a special function (such as indicating titles, headings, or some semantic emphasis) beyond normal phonetic use.

Composite and Ligature Forms

Some EVA "letters" actually represent complex Voynich glyphs or common ligatures. For instance, the sequences ch and sh in EVA correspond to single glyphs often called "bench" characters (shaped like two connected curves with a bench-like top). These bench glyphs are quite frequent, typically at word beginnings (e.g. in chedy, shol – common words discussed below). Similarly, EVA uses double letters like ee or ii to indicate two identical strokes, which in the script often appear as one unit (e.g. a lengthened stroke). We treat these as separate characters for frequency analysis, but note that in the manuscript they may function as digraphs or elongated forms of one sound.

There are also a few rare symbols (e.g. EVA x and q). The glyph transcribed as q is especially noteworthy: it almost always appears in the digraph qo at the start of words. In fact, Voynich q is essentially a word-initial glyph that virtually never occurs without immediately being followed by o – indicating qo acts as a fixed prefix unit in the script.

Table 1: Examples of Voynich Glyphs and Features (EVA transcription)

Symbol Frequency Analysis

A fundamental step is quantifying how often each symbol or combination occurs. Below we present frequency analyses for unigrams (single glyphs), bigrams (pairs of consecutive glyphs), and trigrams (triplets), based on the EVA transcription of the manuscript.

Unigram Frequencies (Single Glyphs)

The Voynich script has a skewed character frequency distribution similar to vowels/consonants in real languages. Table 2 shows the top Voynich characters by frequency:

Table 2: Top 5 Most Frequent Characters (Unigrams)

Counts are estimates based on ~170k total characters and percentage frequencies given in research logs. Subsequent characters (not shown) include y, d, s, c, h etc., each making up a few percent of the text.

Common Bigrams (Glyph Pairs)

Analyzing frequently recurring two-glyph sequences provides insight into Voynich phonotactics (which letter combinations are allowed or favored):

• Repeated Strokes: The sequence ii is one of the most frequent bigrams. This doubling occurs in many common words (e.g. daiin, otaiin) whenever the i glyph is repeated. In the script, this often appears as a lengthened stroke. The prevalence of "ii" suggests a sound or letter that often doubles (analogous to how some languages use double letters or a geminate sound), or it could be a diacritic elongation in the script.
• Vowel-Consonant Combos: ai is another ubiquitous pair, often spanning a transition from a loop glyph a to a stroke i. It appears in the sequence "-aiin" found at many word endings (discussed below). Similarly, oi and od (if we consider d the bench loop) occur frequently in middle positions of words (e.g. qokeedy contains …ko + ee + dy, yielding ke, ed, dy bigrams). Sequences involving o followed by another letter are common, since o often acts as a prefix or connecting vowel.
• Bench + Following Letter: The bigrams ch and sh stand out, but recall each of these represents a single bench glyph, often followed by another letter. If we consider the actual glyph pairs in the manuscript, a bench glyph followed by a vowel is very common (e.g., ch + e in chedy, or sh + o in shol). In EVA bigram terms, ch (c followed by h) occurs frequently at the start of words (since many words begin with the bench glyph for ch). Likewise sh (s followed by h) is a frequent word-initial combination. These reflect that bench characters initiate a large fraction of words, a structural trait of Voynichese.
• Suffix Bigrams: A number of highly frequent bigrams correspond to common endings of words. For example, dy occurs in many word finals (such as -edy in chedy, shedy, or -eedy in keedy, teedy). Similarly in is frequent as part of the ending -iin (in daiin, otaiin, etc.), and ar or or appear when words end in an r-like curve. The ed sequence often appears just before a final y or y-like glyph (-edy, -edy), making ed a common penultimate bigram. These prevalent pairs hint at morphological suffixes – we see the same endings reused across different words (e.g. -dy, -iin), implying a functional ending or grammatical marker.

To illustrate, here are a few of the most common bigrams identified:

• ii (double i stroke) – appears in many plural or extended endings.
• ch (bench) – frequent word-initial bigram (bench glyph beginning).
• sh (bench with loop) – another common word-initial bigram.
• ed – often part of -ed/-edy endings in process words.
• dy – a very common word-final pair in various contexts.
• ai – occurs in the middle and end of many words (especially before n in -ain/-aiin).
• qo – the quintessential prefix; qo- starts many words (so q and o co-occur as a bigram by rule).

These recurring pairs reinforce the idea that Voynich words have stable internal structure – certain combinations are consistently used, much like bigrams "qu" or "th" in English.

Common Trigrams (Three-Letter Sequences)

Frequent three-glyph sequences shed light on typical syllable or morpheme structures in the Voynich text. Some of the top trigrams include:

• iin – This trigram appears extremely often, as it spans the ending -iin found in a host of common words (e.g., daiin, otaiin, okaiin, lkaiin). The sequence iin (often rendered as an i followed by a benched iin shape) could represent a recurring suffix or inflection. It is so frequent that it likely conveys a general grammatical ending (possibly nominal, e.g. akin to a case or plural ending). Even viewed purely structurally, the consistency of -iin at word ends is one of the strongest patterns in the manuscript.
• aii – Another variant of the above, capturing the preceding letter as well. In words ending in -aiin, the trigrams include both aii and iin. The presence of aii (as in "d-aii-n") is essentially the vowel + doubled-i combination near the end of many words. This again underscores how a + ii often functions as part of a larger suffix.
• edy – This trigram spans the common ending -edy (or -eedy if we include double e) found in words like chedy, shedy and their extended forms (keedy, teedy, etc.). -edy appears to be a productive suffix, likely indicating some process or quality. As a trigram, edy is highly frequent and nearly always occurs at word finals. Similarly, eey (for -eey as in okeey) and ody (as in qokody if it existed) might show up, but edy is particularly prominent due to multiple high-frequency words sharing it.
• che – The trigram covering a bench + e sequence. For instance, chedy begins with che, and cheol (if present) would as well. This indicates a bench glyph followed by e is a frequent syllable onset (che-). Likewise sho (as in shol) could be common, combining the other bench with an o vowel.
• Other structural trigrams: qok (found at the start of qokeedy and qokain) appears often in the astronomical context where qo is prefixed to words starting with k. ain is another notable trigram: it forms the ending of words like dain (a standalone word) and appears inside longer ones (otain part of otaiin). In the context of the manuscript, sequences ending in ain or aiin are so widespread that multiple distinct words share those as endings, reinforcing an agglutinative structure where a root plus a common ending form different words.

In summary, Voynich trigrams often correspond to word stems plus affixes. The consistency of certain triples across many words suggests that if we treat the text as encoded language, these trigrams could map to recurring morphemes (e.g. -iin as a noun ending, -edy as an abstract noun or participle ending, etc.). Importantly, the trigram frequencies again highlight that the script is not random: some triples are omnipresent while others never occur, indicating clear phonotactic rules.

Positional Distribution: Line-Initial, Medial, and Line-Final Patterns

Beyond raw frequencies, Voynich glyphs show striking positional behaviors – certain symbols or words prefer the beginnings or ends of lines, or specific locations within words:

In summary, glyph distribution by position indicates that the Voynich script likely follows a set of orthographic rules. Certain glyphs serve as preferred starters (line-initial or word-initial capitals?), others as preferred finishers, and some only in medial roles. Recognizing these patterns is crucial: it means any proposed decipherment must respect these positional behaviors (just as a valid English plaintext wouldn't start a word with "ng-", a valid Voynich solution must explain why, say, q only ever appears as qo- at the front, or why daiin so often caps a line).

Frequent Word Forms and Recurrent Patterns

Using EVA transcription, we can identify entire words (space-delimited sequences) that recur throughout the manuscript. Table 3 below lists some of the most frequent Voynich words and their raw occurrence counts, along with notes on where they predominantly appear:

Table 3: High-Frequency Voynich Words and Their Distribution

Table 3 above is based on combined analysis of the corpus logs and lexicon data. Counts are total occurrences in the manuscript. Not all high-frequency words are shown; we focused on those over ~200 occurrences.

Key Observations from Frequent Words

• Common Suffixes: Many of the top words share endings. For example, daiin, otaiin, okaiin, lkaiin all end in -aiin. Likewise, chedy, shedy, qokeedy, keedy, teedy end in -edy/-eedy. This reinforces that Voynich words seem to be built from stems + affixes. The -aiin ending is ubiquitous for what appear to be nouns or noun-like terms (especially botanical ones), whereas -edy (or extended -eedy) shows up in many process or descriptive terms (perhaps analogous to an adjective or participle ending). The frequency of these patterns suggests an agglutinative structure, where a root word is modified by adding standard suffixes to convey different meanings or functions.
• Section-Specific Vocabulary: The distribution of frequent words is not uniform across the manuscript's sections. Instead, certain high-frequency terms cluster in specific sections, implying topic-specific vocabulary.

• Formulaic Phrases: Many of these high-frequency words co-occur in repeated phrases or constructs. For example, a pattern like "otaiin shedy qokeedy daiin chedy …" (which is actually a sequence found on a page) shows multiple frequent words chained together. This hints that the text follows a recipe-like or instructional formula. We can imagine a template such as: {Ingredient/Part} + {preparation/process term} + {another ingredient/medium} + {action term}. The recurrence of the same terms in predictable slots supports this idea. In fact, analysts in later phases deduced an underlying pattern akin to "OBJECT – QUALIFIER – PROCESS – AGENT", consistent with how recipes or prescriptions are written.
• Core Vocabulary vs. Rare Words: The frequent words listed constitute a core vocabulary that covers a large portion of the manuscript's content (one analysis showed the top ~50 words account for two-thirds of all tokens). Outside of this core, there are hundreds of rarer words (many appearing only once or a few times). These may include specific plant names or unique references. The existence of a core repeating lexicon alongside a long tail of unique terms is yet another hallmark of a real language or a content-rich cipher.

In summary, by extracting the most frequent word forms, we see clear evidence of structure and topical differentiation in the Voynich manuscript. Each section employs its own subset of vocabulary, yet all sections share certain structural suffixes and prefixes. This suggests a single underlying language or coding system with specialized jargon for different domains (herbal, astrological, medicinal).

Sectional Variations in Glyph Sequences

As touched on above, the Voynich Manuscript's content divisions (sections) correspond to shifts in writing patterns. Here we detail how glyph and word usage varies by section, underscoring that any decipherment hypothesis must account for these contextual variations:

Botanical Section (Herbal Folios)

Each herbal folio typically features an illustration of a plant, a label or name near the plant, and a paragraph of text (presumed to describe the plant's properties or usage). In this section, plant-part terms dominate the text. Words ending in -aiin are especially frequent – e.g., otaiin ("leaf?") and okaiin ("flower?") occur on many pages alongside daiin ("root/seed?"). A typical line in a botanical paragraph might mention giving or preparing the daiin of a plant with its otaiin, etc. The consistent presence of daiin near illustrations of roots is a strong clue: it appears dozens of times specifically in captions or lines adjacent to plant root drawings. This positional concordance suggests daiin is contextually linked to roots or foundational elements.

Structurally, the botanical text often follows a pattern: Plant Name – part/ingredient – action. Glyph-wise, we see many bench-initial words (ch, sh) describing processes, but also a high occurrence of straightforward noun forms (lots of -iin endings without extra suffixes, perhaps denoting ingredients). The overall impression is that the botanical section's writing is somewhat more repetitive and formulaic, likely because each entry follows a template (identifying parts, uses, and preparations for each plant).

Astronomical & Astrological Section

This section includes circular diagrams of zodiac symbols, cosmological charts, and what appear to be star names or month names around the zodiac medallions. The text here is often arranged in radial fashion or as labels around diagrams. Voynich glyph sequences in this section differ from the herbal text in a few ways. First, qo-prefixed words are very prominent (far more than in other sections). For example, qokeedy is heavily used around the star illustrations, and other terms starting with qo or qo* occur near zodiac symbols. The prefix qo- itself might indicate a classification (the logs hypothesized it could mean something like "celestial" or an honorific).

Additionally, this section may contain unique bigrams or names that don't appear elsewhere – possibly transliterations of star names or specialized terminology. The structure of phrases here might incorporate numbers or dates as well (there are instances of sequences that look like they could be ordinal indicators). Interestingly, the astronomical pages still use common words like shedy or daiin occasionally, but in different contexts (perhaps linking celestial events to herbal recipes). For instance, one might see otaiin qoky (as in the lexicon snippet meaning "evening star") – a combination of a plant term with a celestial marker, which is unique to mixed contexts. The presence of such combinations underscores the integrative nature of Voynich content.

Biological/Balneological Section

The so-called "biological" section (folios with naked figures in pools connected by pipe-like structures) introduces a noticeably different lexicon and perhaps even grammar. As Phase 1 analysis found, this section uses less of the botanical vocabulary like daiin, and instead we find more of the process-oriented terms such as chedy and shedy. The texts accompanying these illustrations talk about flows, baths, and possibly bodily functions, so it makes sense that "preparation" or "treatment" words (shedy roughly meaning "prepared" in context, chedy something like "extracted" or "flowing") appear very often. Indeed, shedy is almost unique to this section – it spikes in frequency here compared to others.

The prefix cth (EVA) is a hallmark of this part of the manuscript. Words starting with cth (a combination of the bench c + t + h or a similar compounded shape) occur repeatedly, whereas they are virtually absent from botanical and pharma sections. This suggests new concepts are being discussed – perhaps anatomy or specific treatments – that required extending the script's vocabulary with cth-words. One could speculate cth might be a prefix meaning "body" or "bath" given the context (the log's hypothesis of Greek "chthonic" meaning earth/underworld for body is one idea). Additionally, the biological section often describes sequences of actions (the flow of liquids through the depicted pipes, perhaps), so we see strings of process words.

Pharmaceutical/Recipes Section

This part (overlapping with the latter folios, often featuring drawings of herbs and jars) is where the text reads most like lists and instructions. Glyph sequence analysis here reveals a mixture of the patterns above: bench-initial words, gallows initial words, and lots of suffix repetition. A recipe paragraph might look like a series of phrases: daiin chol qokeedy shol… etc. We see ingredient words (often ending in -iin or -ar) followed by action words (often ending in -edy or -ol).

The pharmaceutical texts have the highest concentration of words like chol, shol, dain, qokain in close proximity, forming a kind of formulaic code. Indeed, by Phase 1's end, a hypothesis emerged that the text follows a universal formula pattern: Authority/Subject + Resource + Quantity + Action. In these recipes, Authority might be an implicit subject (perhaps omitted or indicated by line-initial gallows), Resource corresponds to the ingredient (e.g. a plant part word like daiin), Quantity could be indicated by repeated suffixes or number-glyphs (the manuscript does have peculiar sequences like qo, cho, sho… that may encode numbers), and Action is the process verb (like chedy "extract" or shol "purify").

To summarize the section-wise analysis: each section of the Voynich manuscript uses the core glyph patterns in distinct ways, emphasizing different subsets of the vocabulary. Botanical pages repeat plant-part terms, astrological pages repeat celestial-prefixed terms, biological pages introduce new prefixes for body/bath concepts, and recipe pages mix it all in procedural formulas. Crucially, however, the underlying glyph system remains the same across the manuscript, which suggests a single author or authors sharing a common script and language, with adaptations for subject matter.

Comparative Glyph Structure Mapping with Other Scripts

Although the Voynich script appears unique, Phase 1 analysis benefits from comparing its structural features with those of known ancient scripts and languages. Using the uploaded lexicon datasets and cross-references, we identified several intriguing parallels in glyph usage and combination patterns. These comparative observations provide context – they do not prove any direct relationship, but they help us understand what kind of system Voynichese might be:

Latin Alphabet Influence

Superficially, many Voynich glyphs look like Latin letters. We noted that o, a, e, n, c, i in Voynich resemble their Latin counterparts. This is not merely aesthetic – it could imply the creator of the Voynich script repurposed familiar letter forms to encode a different language or cipher. For example, EVA o is a circle just like Latin "o", and e has a similar loop shape. The overall set of symbols (an alphabet of ~20–25 characters) is also comparable to an alphabetic script. This suggests Voynichese could be an alphabetic writing system (each glyph roughly representing a sound or cipher unit) rather than a logographic system.

However, the word structure (discussed above) does not match any straightforward European language. Despite Latin-looking letters, attempts to read Voynich text as Latin, Italian, Germanic, etc., have all failed. This discrepancy hints that the script might be a cipher: Latin letters (or variants) used to obscure a message in another language or in a coded form. Notably, the gallows characters have no direct Latin equivalent, but they do resemble elaborate script letters or abbreviation symbols used by medieval scribes. Their presence and form (almost like ornate capital letters) further enforce the idea of a medieval European origin.

Resemblance to Hebrew/Arabic Scripts

Some Voynich characters and behaviors bear a passing resemblance to Semitic scripts. For instance, one Voynich glyph looks a bit like Hebrew "aleph" or Arabic "ayn", and the manuscript's use of mostly one-case letters (no uppercase/lowercase distinction) is similar to Hebrew/Arabic. These similarities led researchers to test if Voynich could be written right-to-left like Hebrew. However, as mentioned, reading it R→L did not yield any obvious improvement in pattern recognition or intelligibility.

Still, we see structural analogues: Hebrew, for example, has certain letters that cannot appear at the end of a word (so-called final forms are used instead), akin to Voynich having positional constraints on glyphs. Arabic uses diacritics and letter shape changes depending on position – Voynich bench characters and line-initial flourishes could be an analogous feature. There was also an early idea that maybe Voynich encodes an Arabic or Hebrew text with a special alphabet, but Phase 1 found no straightforward mapping. The main takeaway is that Voynich's creators might have been aware of non-Latin scripts; for example, the concept of an initial marker (like qo-) is reminiscent of how Hebrew often prefixes a letter (vav or yod) to indicate "and" or other grammar, and how Arabic uses the al- article.

"Qo" as a Divine or Authority Marker

A striking cross-script parallel emerged with the qo prefix. In Voynich, qo- begins many high-level terms (especially in astrology and recipes). This is structurally similar to how some ancient scripts prepend special signs to important names or concepts. For instance, in Sumerian cuneiform, the DINGIR sign (a star symbol) was prefixed to divine names or heavenly objects – conceptually marking "this is divine/celestial". Voynich qo appears to function in a comparable way (it even appears heavily in the star-related section).

The Phase 1 enhanced analysis explicitly noted 'qo-' = authority/divine/above in pattern matching, aligning it with not only Sumerian but also Egyptian (the ntr "god" hieroglyph used before deity names), and with divine name markers in scripts from Linear A to Mayan (Maya, for example, often prefixed a glyph for "lord" or "sky" in front of names of kings or sky objects). The fact that 37 out of 41 reference scripts showed a similar concept of an authority or celestial marker gives confidence that qo in Voynich is not random.

"Daiin" and the Universal "Root" Word

The word daiin is the most frequent in Voynich and, as we saw, contextually linked to plant roots or base substances. It is fascinating that virtually every ancient medicinal or administrative tradition has a commonly used word for "base" or "root" that appears constantly in texts. For example, in Akkadian (Mesopotamian herbals) the word for "root" (šeršu) shows up in almost every recipe since many medicines use roots. In Egyptian, "bnr" (root) is common in medical papyri.

The Voynich daiin might be playing that same role – structurally, it's a short, frequently repeated noun that could mean "root, foundation, ingredient." What's more, daiin appears in patterns that match those languages: often following a plant name or preceding a preparation verb, exactly as "root" would in a recipe instruction (e.g., "take [plant] root and…" is a formula in many traditions). The lexicon comparisons show words for root or seed in dozens of languages, and daiin aligns with that semantic field across 39 out of 41 script traditions checked.

Repetitive Suffixes and Plural/Case Markers

Voynich's habit of repeating characters (like -eedy vs -edy, or -aiin with double i) has analogues in other writing systems. Egyptian hieroglyphs famously indicate plural by repeating a sign three times (or adding stroke marks) – a kind of visual duplication to mean a quantitative change. In Voynich, we see something possibly comparable: a single -edy ending versus a double-e -eedy ending could indicate a modified meaning (plural, intensive, or a different grammatical case). Likewise, doubling ii might indicate a longer sound or plural.

Sumerian and other cuneiform languages used numerical classifiers or repeated signs to denote measures and plurality. Even Linear B (Mycenaean Greek) had separate signs or repeated symbols for plural. The Voynich data shows a systematic pattern of repetition: when an ending is important, it might be extended. For example, shedy vs sheedy (the latter appears rarely but could be an intensified form of the former in theory). The cross-script perspective suggests we should interpret these not as mistakes or random, but as purposeful markers. The Phase 1 comparative analysis noted that such quantity or grammatical patterns in Voynich "match… Egyptian plural strokes, Linear B counting systems, [and] Mesoamerican bar-dot variations".

Formulaic Text Structures

The content format of Voynich (especially the recipes) was found to mirror a universal administrative formula: Authority + Resource + Quantity + Action. Many ancient documents, from Babylonian inventories to medieval Latin recipes, follow a set order of information. For example, a typical formula might be "Take X of ingredient Y and do Z." Voynich lines exhibit a similar consistent ordering of elements.

The comparative study with 41 script traditions revealed that such patterns are nearly ubiquitous in historical texts of certain genres (100% of checked Mediterranean administrative texts had an Authority-Resource-Quantity-Action pattern, 95% of Near Eastern trade documents, etc.). Voynich's text, by aligning with this pattern, reinforces that it likely encodes practical information (not pure cipher gibberish).

In conclusion, the comparative glyph and pattern mapping situates the Voynich Manuscript in a broader context of human writing systems. While Voynichese's specific symbols are unique, the roles those symbols play are not unheard of. Authority markers, base resource words, plural indicators, and formulaic constructions are all well-attested in historical texts – and Voynich exhibits all of them in ciphered form. This strongly supports the idea that Voynichese is not a hoax or unsystematic invention but a purposeful encoding of meaningful text, likely using a cipher or constructed language drawing inspiration from multiple scripts.

Conclusion (Phase 1 Findings)

"Phase 1 has brought the Voynich Manuscript out of the realm of pure mystery and into a structured framework – we can now say with confidence that the manuscript's text is highly patterned and likely purposeful."